Measuring quality of life in patients with sleep apnoea: whose life is it anyway?
http://www.100md.com
《胸》
Correspondence to:
Dr W W Flemons
Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada T2N 2T9; flemons@ucalgary.ca
A new self-administered disease specific questionnaire for sleep apnoea
Keywords: obstructive sleep apnoea; quality of life; questionnaire; Quebec Sleep Questionnaire
Patients suspected of suffering from sleep apnoea present to primary care physicians with several typical symptoms
including habitual snoring (often disruptive to bed partners), waking feeling unrefreshed, daytime sleepiness, or
fatigue. The disorder is increasingly being recognised in patients who are hypertensive, obese, or who have
unexplained respiratory failure. Randomised placebo controlled trials have proved that sleep apnoea has an adverse
impact on mood, quality of life, functional status, and vigilance, and that treatment with continuous positive
airway pressure (CPAP) results in statistically significant improvements. Although sleep apnoea is strongly
suspected to be a risk factor for developing systemic hypertension, and some preliminary evidence suggests that it
is associated with an increased risk for cardiovascular and cerebrovascular disease, there is no convincing
evidence yet that CPAP reduces these risks.1 In most patients treatment is therefore primarily aimed at improving
their quality of life. Although symptoms and quality of life are often used interchangeably, the former is a subset
of the latter and does not capture the complete impact of the disorder.
When patients are treated for sleep apnoea it is important to document whether the treatment is effective. All too
often this evaluation is limited to determining whether the apnoea-hypopnoea index has been satisfactorily reduced.
A Medline search of the English language literature for randomised controlled trials of adult sleep apnoea
syndromes over the past 10 years produced 95 studies, 40% of which did not report any symptom or quality of life
outcomes. It is well recognised that the apnoea-hypopnoea index correlates poorly with these outcomes, so by itself
the index is not an appropriate measure. If it is accepted that quality of life is what matters most to patients
with sleep apnoea, then clinical trials should include it as an important and possibly the primary outcome measure.
Quality of life was included as an outcome measure in only 25% of the 95 clinical trials.
When an investigator is planning to conduct a clinical trial or when a practitioner is reviewing a published study,
what expectations should they have for the method by which quality of life is measured? In general there are two
categories of quality of life measures—generic and disease specific.2 Generic indices such as the SF-36 have the
advantage that they are in common use and have been used in many disease states so it is possible to compare
impacts of sleep apnoea and its treatment with other disorders. However, the drawback to these indices is that they
may fail to capture important aspects of the impact of sleep apnoea on patients and therefore may be insensitive to
important improvements experienced with treatment. In addition, generic indices were designed to compare broad
aspects of quality of life at a single point in time across disease states (a discriminative property) and were not
specifically designed to measure within-subject change following a therapeutic intervention (an evaluative
property).2 Quality of life indices, like physiological measures, should be evaluated for their signal to noise
properties before being accepted as important measures in clinical trials. The "signal" in a subjective measure
like quality of life is the magnitude of the change in score when patients have improved or deteriorated, a
property referred to as responsiveness. The "noise" in this type of measure is the amount of change recorded in
stable patients who have not yet been offered treatment, a property referred to as reliability. Most symptom
questionnaires are not rigorously evaluated for these two essential properties or for a third property referred to
as validity. An instrument is considered valid when it has been shown to measure what it is intended to.
Disease specific questionnaires such as the Quebec Sleep Questionnaire (QSQ) for sleep apnoea described by Lacasse
et al3 in this issue of Thorax are developed by a long and arduous process that begins with patients’ opinions
about the most frequent and important aspects of the disorder that impact on their lives. This ensures face
validity of the instrument and, in the case of the QSQ, the fact that its items are very close to the similarly
designed sleep apnoea quality of life index4,5 adds to its face validity. The QSQ was shown to have construct
validation by demonstrating that within-subject changes in its domains correlate, as predicted, with other quality
of life (SF-36; Symptom Checklist-90), functional status (Functional Outcomes in Sleep Questionnaire), and symptom
based questionnaires (Epworth Sleepiness Scale). So why not simply use these instruments? Why another one? The QSQ
was designed as an evaluative instrument, meaning that the authors included items that patients indicated during
the construction phase of the questionnaire would be sensitive to change with treatment. It is therefore more
likely to detect small but important changes in the lives of patients with sleep apnoea (signal) than the other
instruments. In fact, the authors were able to show that it has a stronger signal than the Functional Outcomes in
Sleep Questionnaire. Importantly, it has also been shown to have a relatively small amount of noise (high
reliability), as evidenced by its high intraclass correlation coefficient in subjects who completed the
questionnaire twice before being offered treatment. It therefore has excellent signal to noise ratio properties.
Perhaps the most important property of a quality of life questionnaire used in clinical trials is the ability of
its results to be understood or placed in context. It is also the property most overlooked in quality of life or
symptom based questionnaires. It answers the question appropriately asked about many results—"So what?" If a study
shows that four out of the eight domains of the SF-36 improve significantly (statistically), what is the clinical
significance of this? If one domain improves by 25%, is this a large increase? Does this magnitude of change
indicate a life changing event or merely a minor improvement in one aspect of a patient’s life? This property of a
questionnaire is referred to as interpretability and should be grounded in the experience of patients, not in
statistics.2 No generic quality of life instrument or symptom based questionnaires have published evidence on
interpretability. Only nine of the published sleep apnoea clinical trials (<10%) used a disease specific quality of
life instrument and most of these had not described the property of interpretability. It is therefore difficult, if
not impossible, to estimate from the published literature the clinical importance and magnitude of the change in
patients’ lives when they have been adequately treated for sleep apnoea.
The QSQ, like the Sleep Apnoea Quality of Life Index, provides "consumers" with a minimal important clinical
difference. This draws a line in the sand which allows researchers to describe the percentage of patients improved
to this extent, and suggests that a statistically significant mean result that is less than this may not be
clinically significant. The QSQ has evaluated this property in a relatively small number of patients and obtained
rather large numbers for their minimal important clinical difference—much larger than similarly constructed
questionnaires for other disease states and for the Sleep Apnoea Quality of Life Index.4 The exact reason for this
discrepancy is not clear from the description of their methods or the discussion of their results. It may be
related to the small numbers of patients studied or because they chose to use a self-administered format rather
than the interviewer administered format of most other questionnaires. Additional research is required to evaluate
this property further. Notwithstanding this, the development of the QSQ establishes an important standard for a
self-administered sleep apnoea disease specific questionnaire to which researchers and readers of clinical trials
in this field should pay attention.
REFERENCES
Flemons WW. Clinical practice. Obstructive sleep apnea. N Engl J Med 2002;347:498–504.
Guyatt GH, Feeny DH, Patrick DL. Measuring health-related quality of life. Ann Intern Med 1993;118:622–9.
Lacasse Y, Bureau M-B, Series F. A new standardised and self-administered quality of life questionnaire specific to
obstructive sleep apnea. Thorax 2004;59:494–9.
Flemons WW, Reimer MA. Measurement properties of the Calgary sleep apnea quality of life index. Am J Respir Crit
Care Med 2002;165:159–64.
Lacasse Y, Godbout C, Series F. Independent validation of the Sleep Apnoea Quality of Life Index. Thorax
2002;57:483–8.(W W Flemons)
Dr W W Flemons
Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada T2N 2T9; flemons@ucalgary.ca
A new self-administered disease specific questionnaire for sleep apnoea
Keywords: obstructive sleep apnoea; quality of life; questionnaire; Quebec Sleep Questionnaire
Patients suspected of suffering from sleep apnoea present to primary care physicians with several typical symptoms
including habitual snoring (often disruptive to bed partners), waking feeling unrefreshed, daytime sleepiness, or
fatigue. The disorder is increasingly being recognised in patients who are hypertensive, obese, or who have
unexplained respiratory failure. Randomised placebo controlled trials have proved that sleep apnoea has an adverse
impact on mood, quality of life, functional status, and vigilance, and that treatment with continuous positive
airway pressure (CPAP) results in statistically significant improvements. Although sleep apnoea is strongly
suspected to be a risk factor for developing systemic hypertension, and some preliminary evidence suggests that it
is associated with an increased risk for cardiovascular and cerebrovascular disease, there is no convincing
evidence yet that CPAP reduces these risks.1 In most patients treatment is therefore primarily aimed at improving
their quality of life. Although symptoms and quality of life are often used interchangeably, the former is a subset
of the latter and does not capture the complete impact of the disorder.
When patients are treated for sleep apnoea it is important to document whether the treatment is effective. All too
often this evaluation is limited to determining whether the apnoea-hypopnoea index has been satisfactorily reduced.
A Medline search of the English language literature for randomised controlled trials of adult sleep apnoea
syndromes over the past 10 years produced 95 studies, 40% of which did not report any symptom or quality of life
outcomes. It is well recognised that the apnoea-hypopnoea index correlates poorly with these outcomes, so by itself
the index is not an appropriate measure. If it is accepted that quality of life is what matters most to patients
with sleep apnoea, then clinical trials should include it as an important and possibly the primary outcome measure.
Quality of life was included as an outcome measure in only 25% of the 95 clinical trials.
When an investigator is planning to conduct a clinical trial or when a practitioner is reviewing a published study,
what expectations should they have for the method by which quality of life is measured? In general there are two
categories of quality of life measures—generic and disease specific.2 Generic indices such as the SF-36 have the
advantage that they are in common use and have been used in many disease states so it is possible to compare
impacts of sleep apnoea and its treatment with other disorders. However, the drawback to these indices is that they
may fail to capture important aspects of the impact of sleep apnoea on patients and therefore may be insensitive to
important improvements experienced with treatment. In addition, generic indices were designed to compare broad
aspects of quality of life at a single point in time across disease states (a discriminative property) and were not
specifically designed to measure within-subject change following a therapeutic intervention (an evaluative
property).2 Quality of life indices, like physiological measures, should be evaluated for their signal to noise
properties before being accepted as important measures in clinical trials. The "signal" in a subjective measure
like quality of life is the magnitude of the change in score when patients have improved or deteriorated, a
property referred to as responsiveness. The "noise" in this type of measure is the amount of change recorded in
stable patients who have not yet been offered treatment, a property referred to as reliability. Most symptom
questionnaires are not rigorously evaluated for these two essential properties or for a third property referred to
as validity. An instrument is considered valid when it has been shown to measure what it is intended to.
Disease specific questionnaires such as the Quebec Sleep Questionnaire (QSQ) for sleep apnoea described by Lacasse
et al3 in this issue of Thorax are developed by a long and arduous process that begins with patients’ opinions
about the most frequent and important aspects of the disorder that impact on their lives. This ensures face
validity of the instrument and, in the case of the QSQ, the fact that its items are very close to the similarly
designed sleep apnoea quality of life index4,5 adds to its face validity. The QSQ was shown to have construct
validation by demonstrating that within-subject changes in its domains correlate, as predicted, with other quality
of life (SF-36; Symptom Checklist-90), functional status (Functional Outcomes in Sleep Questionnaire), and symptom
based questionnaires (Epworth Sleepiness Scale). So why not simply use these instruments? Why another one? The QSQ
was designed as an evaluative instrument, meaning that the authors included items that patients indicated during
the construction phase of the questionnaire would be sensitive to change with treatment. It is therefore more
likely to detect small but important changes in the lives of patients with sleep apnoea (signal) than the other
instruments. In fact, the authors were able to show that it has a stronger signal than the Functional Outcomes in
Sleep Questionnaire. Importantly, it has also been shown to have a relatively small amount of noise (high
reliability), as evidenced by its high intraclass correlation coefficient in subjects who completed the
questionnaire twice before being offered treatment. It therefore has excellent signal to noise ratio properties.
Perhaps the most important property of a quality of life questionnaire used in clinical trials is the ability of
its results to be understood or placed in context. It is also the property most overlooked in quality of life or
symptom based questionnaires. It answers the question appropriately asked about many results—"So what?" If a study
shows that four out of the eight domains of the SF-36 improve significantly (statistically), what is the clinical
significance of this? If one domain improves by 25%, is this a large increase? Does this magnitude of change
indicate a life changing event or merely a minor improvement in one aspect of a patient’s life? This property of a
questionnaire is referred to as interpretability and should be grounded in the experience of patients, not in
statistics.2 No generic quality of life instrument or symptom based questionnaires have published evidence on
interpretability. Only nine of the published sleep apnoea clinical trials (<10%) used a disease specific quality of
life instrument and most of these had not described the property of interpretability. It is therefore difficult, if
not impossible, to estimate from the published literature the clinical importance and magnitude of the change in
patients’ lives when they have been adequately treated for sleep apnoea.
The QSQ, like the Sleep Apnoea Quality of Life Index, provides "consumers" with a minimal important clinical
difference. This draws a line in the sand which allows researchers to describe the percentage of patients improved
to this extent, and suggests that a statistically significant mean result that is less than this may not be
clinically significant. The QSQ has evaluated this property in a relatively small number of patients and obtained
rather large numbers for their minimal important clinical difference—much larger than similarly constructed
questionnaires for other disease states and for the Sleep Apnoea Quality of Life Index.4 The exact reason for this
discrepancy is not clear from the description of their methods or the discussion of their results. It may be
related to the small numbers of patients studied or because they chose to use a self-administered format rather
than the interviewer administered format of most other questionnaires. Additional research is required to evaluate
this property further. Notwithstanding this, the development of the QSQ establishes an important standard for a
self-administered sleep apnoea disease specific questionnaire to which researchers and readers of clinical trials
in this field should pay attention.
REFERENCES
Flemons WW. Clinical practice. Obstructive sleep apnea. N Engl J Med 2002;347:498–504.
Guyatt GH, Feeny DH, Patrick DL. Measuring health-related quality of life. Ann Intern Med 1993;118:622–9.
Lacasse Y, Bureau M-B, Series F. A new standardised and self-administered quality of life questionnaire specific to
obstructive sleep apnea. Thorax 2004;59:494–9.
Flemons WW, Reimer MA. Measurement properties of the Calgary sleep apnea quality of life index. Am J Respir Crit
Care Med 2002;165:159–64.
Lacasse Y, Godbout C, Series F. Independent validation of the Sleep Apnoea Quality of Life Index. Thorax
2002;57:483–8.(W W Flemons)