© The Author 2007. Published by Oxford University Press.
Translating the Science of Patient-Reported Outcomes Assessment Into Clinical Practice
Affiliations of author: QOL Consulting, Vancouver, BC, Canada; Department of Medicine, University of British Columbia, Vancouver, BC, Canada
Correspondence to: David Osoba, MD, FRCPC, 4939 Edendale Court, West Vancouver, BC, V7W 3H7 Canada (e-mail: david_osoba{at}telus.net).
 |
ABSTRACT |
|---|
 Top Abstract IntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
Patient-reported outcomes (PROs) are based on direct reporting by patients without the intervention of an observer. They include the self-assessment of functional status, symptoms, and other concerns such as needs and satisfaction with care. Health-related quality of life (HRQOL) assessment is a form of PRO and often includes both functional status and symptoms. The science underlying the assessment of HRQOL in clinical practice requires an understanding of the relationships between symptoms, functional status, and HRQOL, as well as instrument selection, and analysis and interpretation of the data. A modification of the Wilson and Cleary model is proposed to show the likelihood of bidirectional relationships between symptoms, functions, and HRQOL. Instrument selection should be based on the measurement properties of the instruments and patient populations in which they will be used. Analyses of data that allow a calculation of the proportion of patients who benefit from an intervention are preferred to analyses that show only the mean change in scores from baseline. HRQOL assessment in clinical practice has been shown to lead to a better understanding of patients' concerns with improvement in counseling and referral for required services. Potentially, HRQOL assessment should also be used to monitor the progress of a patient's disease and benefit from treatment.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
The past 25 years have seen the gradual introduction of patient-reported outcomes (PROs) in oncology clinical trials. However, the knowledge gained from this experience is just now beginning to be evaluated in daily clinical practice. Implementation in clinical practice has been assisted by the development of technology, including the use of portable computers with touch-sensitive screens and software programs, that allows immediate analysis of the data and makes the results available to health care professionals at the time of the clinic visit. In addition, the exposure of health care professionals to the useful information obtained in clinical trials is likely to have made them increasingly aware of the potential value of assessing PROs in clinical practice. While assessment of PROs in daily practice is now a practical reality, there is still a need for a better understanding of the scientific basis underlying both symptom assessment and health-related quality of life (HRQOL) assessment. This paper explores the scientific basis for construction of appropriate models linking symptoms, functioning, and HRQOL; the choice of measurement instruments; and the analyses and interpretation of the data. While the term HRQOL is used throughout most of the remainder of the paper, the concepts and science are also applicable to other PROs.
|
Patient-Reported Outcomes and Health-Related Quality of Life |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
A PRO is "any report coming directly from a patient (i.e., study subject) about a health condition and its treatment" (1). The distinction between a PRO and an observer-generated report is that the PRO is a self-report made directly by the patient without being made through another party. Thus, toxicity grading of a symptom such as fatigue is not a PRO because the grading was made by someone other than the patient himself/herself. There are several kinds of PROs with the most common being self-reported symptoms and self-reported functioning. Sometimes the symptoms are reported in a specific symptom questionnaire and at other times they are included in a more general HRQOL questionnaire.
|
Relationship Between Symptoms and Health-Related Quality of Life |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
Several questions may be asked about the relationship between symptoms and HRQOL. Do symptoms always precede changes in functioning and HRQOL? Are some symptoms causally related to low HRQOL status while others are the effect of low HRQOL (e.g., do anxiety and depression cause a low global HRQOL or can they be the result of a low global HRQOL)? Are symptoms correlated, e.g., what is the relationship between fatigue and pain, or between insomnia and anxiety? The answers to these questions can elucidate whether there is a difference between symptom assessment and HRQOL assessment and whether the latter is of added value to the former.
A model, based on evidence from relevant data, explaining causality and relationships between symptoms, function, and HRQOL is important for clinical practice because it helps the practitioner to use the most appropriate tools and to correctly interpret the results of HRQOL assessment. In their conceptual model, Wilson and Cleary proposed that there is a unidirectional relationship between several kinds of outcomes, e.g., biological and physiologic phenomena give rise to symptoms (and treatment side effects), which in turn have effects on functioning domains (such as physical, social, and role functioning) (2,3). The constellation of these effects leads to general health perceptions and, ultimately, an individual's concept of his/her overall HRQOL. All of the above are also influenced by innate characteristics and environmental factors. Ferrans et al. (4) have modified the Wilson and Cleary model to make it simpler and have added more complete explanations for the components of the model. Recently, Sousa and Kwok (5) have used structural equation modeling to examine the relationship between the components of the original model. They conclude that the model fits the data (derived from patients with HIV) reasonably well but suggested that links be added between symptoms and general health perceptions and symptoms and HRQOL. The correlations for these links, however, were modest to low. These revisions to the original model show the flow from biological variables through to HRQOL as being unidirectional, although they and Wilson and Cleary have indicated that a flow in the reverse direction might also be possible.
The positioning of the components of the conceptual model with respect to each other appears to have overall validity, as borne out by data summarized by Hahn et al. (6) showing that the correlations between pulmonary physiologic phenomena, measured by high-resolution computerized tomography scans, are higher with FEV1 results (R = .84) than they are with HRQOL scores (R = .33–.40). Clinician-reported minimal capacity exercise is more weakly correlated with HRQOL (R = .57) than is the Medical Research Council dyspnea scale (R = .75).
Given that there are reasonably strong correlations between proximal components of the Wilson and Cleary model and weaker correlations between more distant components, what is the cause-and-effect relationship between symptoms and global HRQOL? The previous models mention, in passing, that the flow of the model may not be strictly unidirectional but do not provide data on this important consideration. If the flow is bidirectional for some of the components, this has important implications for clinical practice. It is axiomatic in practice to treat the cause of a disorder or symptom whenever possible to remove the cause and, thereby, ameliorate its subsequent effects. For example, if anxiety is the cause of low global HRQOL, then the anxiety should be treated, but if anxiety is the result of low global HRQOL (e.g., due to poor physical functioning), then the cause of the poor global HRQOL should be diagnosed and treated.
Fayers and Hand (7) and Fayers et al. (8) contend that certain symptoms are likely to be causally related to global HRQOL but that it is not always clear whether certain other symptoms are causally related or are the effect of low global HRQOL. Thus, while symptoms like nausea and vomiting, diarrhea, dyspnea, and constipation may produce a low global HRQOL, they themselves are unlikely to be the result of a low global HRQOL. These symptoms almost always are evident before there is deterioration of global HRQOL and not the other way around (e.g., nausea and vomiting may precede a low global HRQOL, but a low global HRQOL is unlikely to precede nausea and vomiting). Thus, the above-mentioned symptoms are most likely to be causally related to HRQOL. However, it is more difficult to determine the relationship of some other symptoms, e.g., anxiety, depression, pain, cognitive difficulty, and fatigue, to HRQOL. In a series of patients who had either head and neck cancer or breast cancer when global QOL (item 30 of the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire [EORTC QLQ-C30]) was very poor (score of 1 or 2 out of 7), Fayers and Hand (7) and Fayers et al. (8) found that the relationships between global QOL with the above symptoms as well as with social functioning and cognitive functioning were uncertain. Sometimes they appeared to precede low global QOL but other times they seemed to follow low global QOL.
In a questionnaire such as the QLQ-C30, symptoms and functioning domains are usually easily distinguishable because they are separated from each other as separate domains and symptoms (9). However, even then, two of the five so-called functioning domains, i.e., emotional and cognitive functioning, are also possible symptom domains (based on their item content) and could be viewed as being in a separate category from physical, social, and role functioning. Two of the symptom domains (pain and fatigue) may be either the cause or the effect of global QOL whereas the nausea and vomiting domain is a causal indicator. Most of the single items (dyspnea, diarrhea, constipation) are causal indicators, but insomnia may be both the cause and the effect of low global QOL.
Further evidence for the interrelationships between symptoms and between symptoms and domains of the EORTC QLQ-C30 is provided by a study of 535 patients with a variety of cancers (Table 1) (10). There were moderately strong correlations (Pearson's R > .5) of physical functioning with role and social functioning and global QOL. As would be expected, there was a moderately strong correlation of physical functioning with fatigue. There was a correlation between fatigue and pain, and there were negative correlations of fatigue and pain with global QOL.
|
In some other questionnaires (e.g., the Functional Assessment of Cancer Therapy—General [FACT-G]), it may be impossible to determine if certain domains are causal or effect indicators because each domain is made up of combinations of both symptom and functioning items (11). In these questionnaires, there is usually no set of separately constructed items intended to assess overall (global) HRQOL, so it is not possible to determine if there is a temporal relationship between certain items and overall HRQOL. Many questionnaires rely upon a summation of domain and item scores as a representation of overall HRQOL, but summations of domain scores may not include all the relevant constructs that make up a latent variable like overall HRQOL. An additional concern, even when it is theoretically possible to study temporal relationships, is that the temporal separation between cause and effect may be, in reality, so brief that it is not possible to determine whether a particular indicator, e.g., depression, preceded, was present simultaneously, or followed a decrease in overall QOL. Thus, more research in this area is required, but the preceding data would indicate that the components of the Wilson and Cleary model and its more recent modifications are not unidirectional, at least with respect to the components dealing with symptoms, functioning, general health perceptions, and overall HRQOL. The model might be better represented by showing arrows going in both directions between these components (Fig. 1). This revised model suggests that overall HRQOL may, at times, have an effect on general health perceptions that, in turn, may affect functional status and, thus, affect the expression of certain symptoms. This extension of the previous models is not a great departure from the main themes of these models but a realization that the flow of effects may be bidirectional in some situations.
|
An illustration of the difference between symptom assessment and HRQOL assessment is provided by a study of the effect of antiemetic control in highly emetogenic chemotherapy (12). Eight hundred thirty-two patients with a variety of cancers about to receive initial chemotherapy containing high-dose cisplatin were pretreated with a 5HT3 receptor antagonist antiemetic. They reported the degree to which they experienced nausea and vomiting daily for 5 days after chemotherapy and completed the EORTC QLQ-C30 on the 7th day. On average, all patients experienced fatigue, with the greatest mean change from baseline (–16.9 ± 4.0 on a 0–100 scale) being reported by the group that had both nausea and vomiting and a lesser amount of worsening in those who had nausea alone (–11.7 ± 5.6). Since nausea and vomiting were associated with fatigue, one might expect that there would also be an associated effect on physical functioning. Indeed, the group with both nausea and vomiting reported a mean decrease of 10.0 ± 4.4 in physical functioning while those with only nausea had a decrease of 8.7 ± 5.0 in physical functioning. As might also be expected, there was a worsening in the reported global HRQOL with a mean decrease of 14.2 ± 3.2 in those with both nausea and vomiting and a decrease of 7.4 ± 4.8 with only nausea. In all cases, the differences between those who had nausea and vomiting and those who had no nausea or vomiting were highly statistically significant (P < .001). Thus, there is an apparent relationship between nausea and vomiting, fatigue, physical functioning, and global HRQOL. A study designed to only determine the proportions of patients with control of nausea and vomiting would not have uncovered these relationships and their importance in understanding the broader, and important, effects of the failure to control emesis (13).
|
The Added Value of Health-Related Quality of Life Assessment |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
What else can HRQOL assessment add to assessment of symptoms, even if the study is designed as a symptom control study? In the antiemetic study cited above, an unexpected finding was that the group with no nausea or vomiting also reported a substantial worsening in fatigue (–7.9 ± 5.3) (12). Thus, there were causes of fatigue, perhaps directly related to the chemotherapy, other than just nausea and vomiting in these patients. Had fatigue not been assessed, this finding would have not been discovered. Furthermore, in this study, patients who experienced more than two episodes of vomiting (with nausea) had significantly greater deterioration in mean global QOL scores (P = .003) than did those with one to two episodes of vomiting. This made it clear that the level of global QOL is associated with the frequency of nausea and vomiting. Finally, those with one to two episodes of vomiting had lower global QOL scores than did those with no nausea and vomiting indicating that even one to two episodes of vomiting (in the presence of nausea) had a deleterious effect on global QOL (12). This finding led to the conclusion that the goal of antiemetic treatment for chemotherapy should be complete, and not just partial, control of nausea and vomiting. This goal has been enunciated in antiemetic guidelines (14).
|
What to Measure? |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
The above study shows that measuring the effect of treatment only on symptoms does not provide information about the effect of those symptoms on functioning. Conversely, measuring only functioning will not reveal information about specific symptoms that are of concern to the patient. Measuring both symptoms and functioning, but not global QOL, will not provide information about how these, and other factors, affect a patient's perception of his/her overall (global) HRQOL. Therefore, since it is the patient and not just the symptoms that we should be treating, it is necessary to measure all three components, i.e., symptoms, functioning, and global HRQOL, to obtain as complete a picture as possible of how a disease, its symptoms, and its treatment affect a patient.
|
A Model for Health-Related Quality of Life Assessment in Clinical Practice |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
There is a great deal of information about the value of HRQOL assessment in clinical trials, but the formal collection of HRQOL data is not often used in day-to-day clinical practice. At present there is still little recognition of the potential value of this kind of data. Several barriers to the application of HRQOL assessment in clinical practice have been identified, and most of them have been overcome (15).
One of the important barriers is to understand how HRQOL information might fit into clinical practice. The process of medical decision making in the clinic or office begins when the patient is first seen and the doctor, nurse, and other health care professionals gather information from the patient's medical history and the physical examination. This information is used to establish a list of possible differential diagnoses. Next, laboratory and imaging studies are ordered, and the information from these is used to distinguish between the differential diagnoses to reduce the list or to revise it. Additional tests may be required to confirm the most likely diagnosis. The patient is then treated with appropriate procedures, medications, etc., depending on the most likely diagnosis. As treatment proceeds, further tests may be ordered, physical examinations are repeated, and further questions are asked. Thus, the process is iterative and ongoing, and the diagnosis is revised as necessary until a final diagnosis is established (Fig. 2).
|
Where might HRQOL information fit into the above process? In the current model, HRQOL information from group data in clinical trials is adapted, to the extent possible, to the individual patient. Some of the group data are helpful for counseling patients and choosing the best therapy, but there are wide bounds on the data and they may not be helpful in addressing an individual patient's concerns. An analogy is the use of survival data to counsel an individual. The survival data provide probabilities of survival but cannot accurately predict the length of survival for a particular individual. The desired model for fitting HRQOL data into clinical practice is to use them like a conventional laboratory test, i.e., they should be used to inform the differential diagnosis and also be used for monitoring a patient's progress. If the data can be made available immediately after the patient has completed the questionnaire, by use of touch screen computers (16,17), then they can be used to add to the information gained from the history and physical examination at the initial and subsequent office visits. Indeed, if the patient completes the questionnaire while waiting to see the physician for the first time, this information can be used to inform the functional inquiry that is part of the initial history taking. Completion of questionnaires on subsequent visits can provide data for monitoring the patient's progress. In summary, HRQOL (and PRO) data can be helpful in several phases of a patient's management: during the initial history and physical examination, during laboratory and imaging, during treatment, and during follow-up (Fig. 3). HRQOL information during these phases will assist with making a diagnosis, monitoring the progress of the patient's disease and treatment, and counseling the patient.
|
|
The Modular Approach to Assessing Health-Related Quality of Life in Clinical Practice |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
To choose the best instrument for the purpose intended, it is first important to understand what kind of instruments exist and what they can do. This is analogous to choosing laboratory and imaging procedures since the physician needs to understand what a laboratory test can provide and whether it is applicable in the current clinical situation. Some instruments were designed to be useful in all diseases and in all populations, including healthy ones. These generic instruments assess the most important domains (e.g., physical, emotional, and social) but may include others and also generally have a separate item or items that ask about general health or overall HRQOL. One of the best known examples of a generic questionnaire is the Short Form 36 (18) (Table 2). (In this section and Table 2, it is recognized that there are other instruments in addition to the ones cited, but it is beyond the scope of this paper to provide a comprehensive listing of all available instruments.) Generic questionnaires have been used in surveys of the healthy population and for showing the differences between particular diseases, e.g., diabetes and hypertension, or between a particular disease and the general population. Although they have also been used in clinical trials in oncology, they may not be very sensitive in depicting details of a particular cancer or condition. However, modules of separate items pertaining to a particular disease (disease-specific instruments) may be added to a generic questionnaire to make it more appropriate for a particular cancer.
|
A condition-specific instrument has usually been designed for a particular population or populations of diseases, e.g., cancer or chronic illnesses, and validated in these diseases. They usually have several items pertaining to symptoms organized as separate domains or items, e.g., the EORTC QLQ-C30 (9), or as items interspersed with items pertaining to functioning, e.g., the FACT-G (11). Disease-specific modules can be added as can checklists pertaining to a particular therapy or clinical trial. Some of these condition-specific instruments have an overall health/HRQOL domain (e.g., the QLQ-C30) whereas others do not. These instruments have been used extensively in patients with cancer and have been shown to be valid, reliable, and responsive to change.
With the recent surge in interest in measuring symptoms, a number of instruments have been designed for this purpose. Some symptom-specific measures are designed to work as stand-alone measures of a particular symptom or set of symptoms, e.g., the Functional Living Index—Emesis (19), while others are intended to assess many symptoms, e.g., the M. D. Anderson Symptom Assessment Inventory (20). Others are designed to measure a constellation of symptoms related to a particular set of diseases and are used in conjunction with a generic or condition-specific instrument, e.g., the Brain Cancer Module 20 (21), which was designed to be used together with the QLQ-C30 in patients with brain cancers. A more comprehensive review of the many symptom-specific instruments that are available can be found in the article by Kirkova et al. (22).
Study-specific instruments, generally in the form of checklists of items, can provide very specific information about the effects of a particular therapy, e.g., a chemotherapeutic agent, on a patient (23). They are not intended as stand-alone instrument and are to be added to more general instruments. Since they are usually developed in response to a particular trial and are meant to be used as a list of single items, they do not undergo the rigorous validation of other instruments.
In summary, the choice of instruments is analogous to choosing laboratory tests. It is just as important to choose the most appropriate and sensitive instrument as it is to choose the most appropriate and sensitive laboratory test or imaging procedure for the purpose at hand.
|
Match the Instrument and the Population |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
It is valuable to understand the population of patients in whom HRQOL is to be measured and to choose the correct instrument(s) for each population. Again, the analogy with laboratory tests applies here. For example, in a palliative care setting of a very mixed population with heterogeneous diseases such as cancer, diabetes, etc., a generic instrument plus variable symptom scales for each disease would probably be the most sensible approach. If the population is more homogeneous, e.g., cancer patients, then a condition-specific or disease-specific instrument which contains symptoms scales or to which such scales can be added is reasonable. If the population is quite homogeneous, e.g., stage IV breast cancer overexpressing HER2/neu, then a disease-specific questionnaire to which a trial-specific scale can be added would likely be the best. In each of the above situations, study-specific modules or checklists can add additional important information. The most important determinant behind the choice of instruments is the purpose of the measurement, i.e., the question that will be answered by carrying out HRQOL assessment.
|
Analysis and Interpretation |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
In most clinical trials, the data are analyzed to show whether there is a difference in the mean changes of scores from baseline between the arms of the trial. While this analysis approach can determine whether there are statistically significant changes in scores between treatment groups, it does not allow an interpretation of the clinical significance of the magnitude of the changes. For example, is a difference of 5 points (on a 0–100 scale) between treatment groups meaningful in clinical terms? (24) How much change in scores represents a meaningful improvement or deterioration in HRQOL scores? Several studies on the interpretation of HRQOL data have suggested a similar magnitude of change as having clinical significance. These studies have shown that 0.3–0.5 effect size or standard deviation (25,26), 1 standard error of measurement (27), and 7%–8% of the breadth of the possible scale scores (28,29) represent a meaningful change in scores. These values can be used as cut points to classify patients into improved and deteriorated HRQOL categories. However, since they represent a minimum meaningful change, it is suggested that a change of 10% of the scale breadth be taken as representing a definite change that is perceptible to patients and excludes false "positive" scores (30–32). One or another of the above cut points can be used to determine the numbers of patients whose scores have changed more than the cut point and, hence, the proportions of patients who do, or do not, report improvement after an intervention. This result is more easily interpreted, in clinical terms, than are mean change scores. In addition, if the proportion of patients who have benefited after a treatment is known, then the number of patients that need to be treated for one patient to benefit can be calculated and the cost incurred in achieving benefit or one patient can be ascertained (33).
|
Research Into Measuring Health-Related Quality of Life in Clinical Practice |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
To date, there have been two main groups studying the effects of carrying out HRQOL assessment in daily practice in hospital clinic settings. The Leeds group assessed HRQOL using the QLQ-C30 at baseline and during clinic visits (34). Patients with a variety of cancers completed the EORTC QLQ-C30 and were randomly assigned into three groups. One group completed the QLQ-C30, and their physicians were given the results of the assessment during the clinic visits (intervention group). The second group also completed the QLQ-C30, but the results were not provided to the physicians (attention group). The third group did not complete the questionnaire (control group). At later time points, HRQOL was assessed in all three groups using the FACT-G questionnaire. There were substantial improvements and statistically significant differences between the levels of social well-being and emotional well-being in the intervention group as compared with the control group. Overall, FACT-G scores improved in the intervention group as compared with the control group. Conversely, the control group reported more deterioration in the FACT-G summed score than did the intervention group. The differences between the attention–control group and the intervention group were not as striking but still apparent. The authors concluded that routine repeated assessment generally benefited patients; they reported better HRQOL and emotional functioning. They also found that there was more frequent discussion of chronic nonspecific symptoms (P = .03) during the clinic visit. Thus, HRQOL improvement and improvement in counseling were associated with explicit use of HRQOL data. However, HRQOL considerations played only a minor role in decisions to discontinue or modify palliative treatment.
The Amsterdam group took a different approach (35). Ten physicians and 214 patients were enrolled into the study. The patients completed the QLQ-C30 at three successive outpatient visits, and the physicians were randomly assigned, in a crossover design, to receive or not receive the assessment data. Communication between physician and patient at each clinic visit was assessed by audiotapes. Physician awareness of problems was assessed by comparing patients' ratings on Dartmouth Primary Care Cooperation Information Project/World Organization of National Colleges, Academies and Academic Associations of General Practitioners/Family Physicians charts. It was found that HRQOL-related issues were discussed more frequently in the intervention than in the control group (P = .01). Physicians in the intervention group identified a higher percentage of patients with moderate-to-severe health problems. All physicians and 87% of patients believed that intervention facilitated communication.
|
Monitoring Changes Over Time |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
The suggestion of 10% of the score breadth as being clinically meaningful has been derived from group studies, but how reliable is a 10% change of the scale breadth (or 0.5 standard deviation) for monitoring an individual patient over time? The answer is still unknown because it is untested. To determine if a 10% change represents a meaningful change in clinical practice, it will be necessary to correlate it with other anchors, i.e., to use a change in HRQOL scores as a clinical/laboratory test and to then integrate the change with other clinical variables, such as a change in disease status, a change in a patient's self-reported perception of change, the ability to perform certain functions, or other parameters. HRQOL data could be gathered at consecutive clinic visits, plotted as a continuous record over time, and be used to monitor the changes in scores. Changes that are greater than a preset cut point could alert the health care practitioner to possibly important changes in a patient's HRQOL or disease. Such changes could lead to further questioning and investigations and, when required, to changes in an intervention. In the meantime, changes in PROs can be used as a trigger, or prompt, to delve deeper into possible problems in the areas denoted by individual patients. This can aid in counseling or for referring patients to others for further care.
|
Conclusions |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
Much of the science of HRQOL assessment has not yet been adequately tested in clinical practice, particularly as it applies to individual patients. There are practical feasibility issues to be overcome as well as some of the analysis and interpretation issues discussed above. However, much more research into application of HRQOL assessment in clinical practice can now be done (see suggestions, Table 3) than was possible just a few years ago and the answers to the questions posed will, undoubtedly, be found.
|
|
REFERENCES |
|---|
|
TopAbstractIntroductionPatient-Reported Outcomes and...Relationship Between Symptoms...The Added Value of...What to Measure?A Model for Health-Related...The Modular Approach to...Match the Instrument and...Analysis and InterpretationResearch Into Measuring Health...Monitoring Changes Over TimeConclusionsReferences |
|---|
(1) Burke L, Stifano T, Dawisha S, Food and Drug Administration. Food and Drug Administration draft: guidance for industry sponsored patient-reported outcome measures. Office of New Drugs and the Office of Medical Policy in the Center for Drug Evaluation and Research (CDER) in cooperation with the Center for Biologics Evaluation and Research (CBER) and the Center for Devices and Radiological Health (CDRH) at the Food and Drug Administration. (2006) Rockville (MA): US Dept of Health and Human Services, Food and Drug Administration.
(2) Wilson IB, Cleary PD. Linking clinical variables with health-related quality of life. JAMA (1995) 273:59–65.
(3) Patrick D. Concept of health-related quality of life and of patient-reported outcomes. In: Health-related quality of life and patient-reported outcomes: scientific and useful outcome criteria—Chassany O, Caulin C, eds. (2002) Paris (France): Springer Verlag. 23–34.
(4) Ferrans CE, Zerwic JJ, Wilbur JE, Larson JJ. Conceptual model of health-related quality of life. J Nurs Scholarsh (2005) 37:336–42.[CrossRef][Web of Science][Medline]
(5) Sousa KH, Kwok O-M. Putting Wilson and Cleary to the test: analysis of a HRQOL conceptual model using structural equation modeling. Qual Life Res (2006) 15:725–37.[CrossRef][Web of Science][Medline]
(6) Hahn EA, Cella D, Chassany O, Fairclough DL, Wong GY, Hays RD. Clinical Significance Consensus Meeting Group. A comparison of the precision of health-related quality of life data relative to other clinical measures. Mayo Clin Proc. In press.
(7) Fayers PM, Hand DJ. Factor analysis, causal indicators and quality of life. Qual Life Res (1997) 6:139–50.[Web of Science][Medline]
(8) Fayers PM, Hand DJ, Bjordal K, Groenvold M. Causal indicators in quality of life research. Qual Life Res (1997) 6:393–406.[CrossRef][Web of Science][Medline]
(9) Aaronson N, Ahmedzai S, Bergman B, Bullinger M, Cull A, Duez NJ, et al. The European Organization for Research and Treatment of Cancer QLQ-C30: a quality-of-life instrument for use in international trials in oncology. J Natl Cancer Inst (1993) 85:365–76.
(10) Osoba D, Zee B, Pater J, Warr D, Kaizer L, Latreille J. Psychometric properties and responsiveness of the EORTC Quality of Life Questionnaire (QLQ-C30) in patients with breast, ovarian and lung cancer. Qual Life Res (1994) 3:353–64.[CrossRef][Web of Science][Medline]
(11) Cella DF, Tulsky DS, Gray G, Sarafian B, Linn E, Bonomi A, et al. The functional assessment of cancer therapy scale: development and validation of the general measure. J Clin Oncol (1993) 11:570–9.
(12) Osoba D, Zee B, Warr D, Latreille J, Kaizer L, Pater J. Effect of postchemotherapy nausea and vomiting on health-related quality of life. Support Care Cancer (1997) 5:307–13.[CrossRef][Web of Science][Medline]
(13) Osoba D. Impact of nausea and vomiting on health-related quality of life. In: Management of nausea and vomiting in cancer and cancer treatment—Hesketh PJ, ed. (2005) Sudbury (MA): Jones and Bartlett Publishers, Inc. 99–118.
(14) Gralla RJ, Osoba D, Kris MG, Kirkbride P, Hesketh PJ, Chinnery LW, et al. Recommendations for the use of antiemetics: evidence-based, clinical practice guidelines. J Clin Oncol (1999) 17:2971–94.
(15) Osoba D. Measuring the effect of cancer on quality of life. In: Effect of cancer on quality of life—Osoba D, ed. (1991) Boca Raton (FL): CRC Press Inc. 25–40.
(16) Buxton J, White M, Osoba D. Patients' experiences using a computerized program with a touch-sensitive video monitor for the assessment of health-related quality of life. Qual Life Res (1998) 7:513–9.[CrossRef][Web of Science][Medline]
(17) Velikova G, Brown JM, Smith AB, Selby PJ. Computer-based quality of life questionnaires may contribute to doctor-patient interactions in oncology. Br J Cancer (2002) 26:51–9.
(18) Ware JE, Sherbourne CD. The MOS 36-item short form health survey (SF-36): conceptual framework and items selection. Med Care (1992) 30:473–83.[Web of Science][Medline]
(19) Lindley CM, Bernard S, Fidds SM. Incidence and duration of chemotherapy induced nausea and vomiting in the outpatient oncology population. J Clin Oncol (1989) 7:1142–9.[Abstract]
(20) Cleeland CS, Mendoza TR, Wand XS, Chou C, Harle MT, Morrissey M, et al. Assessing symptom distress in cancer patients: the M.D. Anderson Symptom Inventory. Cancer (2000) 89:1624–46.
(21) Osoba D, Aaronson NK, Muller K, Sneeuw K, Hsu MA, Yung WK, et al. The development and psychometric validation of a brain cancer quality-of-life questionnaire for use in combination with general cancer-specific questionnaires. Qual Life Res (1996) 5:139–50.[CrossRef][Web of Science][Medline]
(22) Kirkova J, Davis MP, Walsh D, Tiernan E, O'Leary N, LeGrand SB, et al. Cancer symptom assessment instruments: a systematic review. J Clin Oncol (2006) 24:1459–73.
(23) Osoba D. Self-rating symptom checklists: a simple method for recording and evaluating symptom control in oncology. Cancer Treat Rev (1993) 19:43–51.[CrossRef][Web of Science][Medline]
(24) Guyatt GH, Osoba D, Wu AW, Wyrwich KW, Norman GR. Clinical Significance Consensus Meeting Group. Methods to explain the clinical significance of health status measures. Mayo Clin Proc (2002) 77:371–83.
(25) Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life: the remarkable universality of half a standard deviation. Med Care (2003) 41:582–92.[CrossRef][Web of Science][Medline]
(26) Farivar SS, Kiu H, Hays RD. Another look at the half standard deviation estimate of the minimally important difference in health-related quality of life scores. Expert Rev Pharmacoeconomics Outcomes Res (2004) 4:521–9.
(27) Wyrwich K, Tierney W, Wolinsky F. Further evidence supporting a SEM-based criterion for identifying meaningful intra-individual changes in health-related quality of life. J Clin Epidemiol (1999) 52:861–73.[CrossRef][Web of Science][Medline]
(28) Jaeschke R, Singer J, Guyatt GH. Measurement of health status. Ascertaining the minimal clinically important difference. Control Clin Trials (1989) 10:407–15.[CrossRef][Web of Science][Medline]
(29) Osoba D, Rodrigues G, Myles J, Zee B, Pater J. Interpreting the significance of changes in health-related quality-of-life-scores. J Clin Oncol (1998) 16:139–144.
(30) Osoba D, Tannock IF, Ernst DS, Neville AJ. Health-related quality of life in men with metastatic prostate cancer treated with prednisone alone or mitoxantrone and prednisone. J Clin Oncol (1999) 17:1654–63.
(31) Osoba D, Brada M, Yung WKA, Prados M. Health-related quality of life in patients treated with temozolomide versus procarbazine for recurrent glioblastoma multiforme. J Clin Oncol (2000) 18:1481–91.
(32) Osoba D, Bezjak A, Brundage M, Zee B, Tu D, Pater J. Analysis and interpretation of health-related quality-of-life data from clinical trials: basic approach of the National Cancer Institute of Canada Clinical Trials Group. Eur J Cancer (2005) 41:280–7.[CrossRef][Web of Science][Medline]
(33) Goldstein RS, Gort EH, Guyatt GH, Feeny D. Economic analysis of respiratory rehabilitation. Chest (1997) 112:370–9.
(34) Velikova G, Booth L, Smith AB, Brown PM, Lynch P, Brown JM, et al. Measuring quality of life in routine oncology practice improves communication and patient well-being: a randomized controlled trial. J Clin Oncol (2004) 22:714–24.
(35) Detmar SB, Muller MJ, Schornagel JH, Wever LD, Aaronson NK. Health-related quality-of-life assessments and patient-physician communication: a randomized controlled trial. J Am Med Assoc (2002) 288:3027–34.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Cirillo, M. Venturini, L. Ciccarelli, F. Coati, O. Bortolami, and G. Verlato Clinician versus nurse symptom reporting using the National Cancer Institute--Common Terminology Criteria for Adverse Events during chemotherapy: results of a comparison based on patient's self-reported questionnaire Ann. Onc., December 1, 2009; 20(12): 1929 - 1935. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. G Contopoulos-Ioannidis, A. Karvouni, I. Kouri, and J. P A Ioannidis Reporting and interpretation of SF-36 outcomes in randomised trials: systematic review BMJ, January 12, 2009; 338(jan12_1): a3006 - a3006. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. A. Ganz and P. J. Goodwin Health-Related Quality of Life Measurement in Symptom Management Trials J Natl Cancer Inst Monographs, October 1, 2007; 2007(37): 47 - 52. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||