Psychometric Performance and Responsiveness of the Functional Outcomes of Sleep Questionnaire and Sleep Apnea Quality of Life Index in a Randomized Trial: The HomePAP Study

Abstract

Study Objectives:

Measures of health-related quality of life (HRQL) specific for sleep disorders have had limited psychometric evaluation in the context of randomized controlled trials (RCTs). We investigated the psychometric properties of the Functional Outcomes of Sleep Questionnaire (FOSQ) and Sleep Apnea Quality of Life Index (SAQLI). We evaluated the FOSQ and SAQLI construct and criterion validity, determined a minimally important difference, and assessed for associations of responsiveness to baseline subject characteristics and continuous positive airway pressure (CPAP) adherence in a RCT population.

Design:

Secondary analysis of data collected in a multisite RCT of home versus laboratory-based diagnosis and treatment of obstructive sleep apnea (HomePAP trial).

Participants:

Individuals enrolled in the HomePAP trial (n = 335).

Interventions:

N/A.

Measurement and Results:

The FOSQ and SAQLI subscores demonstrated high reliability and criterion validity, correlating with Medical Outcomes Study 36-Item Short Form Survey domains. Correlations were weaker with the Epworth Sleepiness Scale (ESS). Both the FOSQ and SAQLI scores improved after 3 mo with CPAP therapy. Averaging 4 h or more of CPAP use was associated with an increase in the FOSQ beyond the minimally important difference. Baseline depressive symptoms and sleepiness predicted FOSQ and SAQLI responsiveness; demographic, objective obstructive sleep apnea (OSA) severity and sleep habits were not predictive in linear regression.

Conclusions:

The Functional Outcomes of Sleep Questionnaire (FOSQ) and Sleep Apnea Quality of Life Index (SAQLI) are responsive to CPAP intervention, with the FOSQ being more sensitive to differences in CPAP adherence than the SAQLI. These instruments provide unique information about health outcomes beyond that provided by changes in physiological measures of OSA severity (apnea-hypopnea index).

Clinical Trial Information:

Portable Monitoring for Diagnosis and Management of Sleep Apnea (HomePAP) URL: http://clinicaltrials.gov/show/NCT00642486. NIH clinical trials registry number: NCT00642486.

Citation:

Billings ME, Rosen CL, Auckley D, Benca R, Foldvary-Schaefer N, Iber C, Zee PC, Redline S, Kapur VK. Psychometric performance and responsiveness of the Functional Outcomes of Sleep Questionnaire and Sleep Apnea Quality of Life Index in a randomized trial: the HomePAP study. SLEEP 2014;37(12):2017-2024.

INTRODUCTION

Disrupted sleep and daytime sleepiness, common symptoms of obstructive sleep apnea (OSA) in adults, are linked to decreased sense of well-being.¹ Subjects with sleep apnea often, although not uniformly, report impaired quality of life, especially in areas of vitality.² The severity of OSA as determined physiologically by the apnea-hypopnea index (AHI) or hypoxemic burden has not typically correlated with generic measures of health-related quality of life (HRQL)³ or sleep-specific HRQL instruments.^4–9 Thus, it is difficult to predict the effect of OSA on patient- reported outcomes based on physiological severity. Well-validated, sleep-specific HRQL instruments are important for evaluating patient reported outcomes. Several instruments have been developed to assess the consequences of poor sleep and daytime sleepiness caused by sleep disorders on role functioning,¹⁰ including the Functional Outcomes of Sleep Questionnaire (FOSQ)¹¹ and the Calgary Sleep Apnea Quality of Life Index (SAQLI).¹² The SAQLI was designed specifically for OSA, whereas the FOSQ was constructed to measure the consequences of sleep disorders in general.

Neither the FOSQ nor the SAQLI have been extensively validated in randomized control trial settings. Initial FOSQ instrument development used a convenience sample of individuals recruited from a sleep clinic visit to demonstrate test-retest reproducibility and internal reliability.¹¹ The FOSQ subscales were subsequently shown to correlate with the Medical Outcomes Study 36-Item Short Form Survey (SF-36)¹³ domains using a sample of 51 OSA subjects participating in a research project.¹¹ The SAQLI has been shown to have high internal reliability. When concurrent validity was assessed using the SF-36 in an initial validation study of 24 patients with OSA, it showed high correlation only with the SF-36 vitality domain.¹² The SAQLI has discriminative ability for OSA: SAQLI scores were higher in those with severe sleep apnea compared to simple snorers.¹² The FOSQ has been widely used in sleep research studies,^6,14–17 but data on its response properties have not been uniformly reported.^4,18 The SAQLI is less widely used; to our knowledge, no further validation of the scale has been done.⁴ Thus, we sought to assess the psychometric properties of the FOSQ and SAQLI in subjects enrolled in an OSA treatment trial and propose a minimally important difference (MID) score change for both instruments.

Continuous positive airway pressure (CPAP) has been shown to improve some measures of HRQL,¹⁹ but does not consistently improve generic HRQL instrument scores.^1,4,18,20 The effect of CPAP treatment on FOSQ scores has been positive,^19,21 but the magnitude of benefit has been variable and at times with small clinical effect.⁴ CPAP adherence may be associated with HRQL changes: greater use of CPAP would be expected to yield greater improvement in HRQL. Two prior studies have examined the dose of CPAP (hours/night) necessary to “normalize” the FOSQ to 18 or higher. A score of less than 18 is considered abnormal, reflecting a negative effect on sleepiness on quality of life. One study identified a threshold of 7 h,²² but this was not seen in a second randomized study of CPAP.²³ In both studies, FOSQ score variations were large and responses were not well correlated with CPAP use. Prior studies did not adjust for differences in baseline factors, which may affect HRQL and responsiveness. Because HRQL could be adversely affected by comorbid insomnia, depression, or limited sleep opportunity, these confounding factors may limit the effect of CPAP therapy on HRQL. Further consideration of these factors may help elucidate the sources of variation in HRQL responses to CPAP therapy in patients with OSA.

The effect of sleep disorders and sleepiness on HRQL is likely mediated by cultural, social, economic, and individual factors. As defined by the World Health Organization, HRQL reflects the individual's perception of his or her life in the context of their culture and value system, in relation to individual goals, expectations, standards, and concerns.²⁴ HRQL is by nature a subjective rating incorporating physical well-being, psychological health, social relationships, and interactions with the environment that produce a sense of gratification with life. As such, we hypothesized that the FOSQ and SAQLI baseline scores and responsiveness to CPAP would be associated with sociodemographics (sex, race, education, and residential socioeconomic status) and sleep opportunity. We also hypothesized that greater CPAP adherence would result in larger instrument response, reflecting an index of “dose” of the intervention, after adjusting for individual factors.

METHODS

We used patient reported outcomes data obtained from a randomized trial, the HomePAP study. As described elsewhere,²⁵ the HomePAP study compared unattended home-based studies to laboratory-based testing for the diagnosis and treatment of sleep apnea in subjects with moderate to severe OSA and a minimum Epworth Sleepiness Scale (ESS)²⁶ score of 12. The primary outcome for the study was CPAP adherence at 3 mo. The study demonstrated noninferiority of the home-based arm compared to the laboratory arm; subjects in the home arm had better adherence overall. Subjects completed questionnaires assessing demographic information, sleep habits as well as the ESS, SF-36,¹³ Center for Epidemiologic Studies Depression Scale (CES-D),²⁷ and FOSQ instruments at enrollment using pen and paper. The SAQLI was administered by the study coordinator at each site. Subjects were randomized to either home or laboratory-based pathways. Subjects verified to have an AHI of ≥ 15 qualified to continue in the trial and were started on CPAP after home or laboratory-based titration. At 3 mo, data on CPAP use were downloaded and subjects completed the same sleep habits questionnaires and instruments.

Instruments

The FOSQ is a 30-item self-administered questionnaire assessing the effect of excessive sleepiness on activities of daily life.¹¹ It includes five subscales identified through factor analysis: activity level, vigilance, intimacy, general productivity, and social outcome.¹ Subjects are asked if sleepiness interferes with performing a given task; responses range from 1 (extreme difficulty) to 4 (no difficulty). Responses are averaged (excluding missing responses) to create a subscale score of 1 to 4, and then subscale scores are summed (5-20 for the total score), with greater scores indicating less effect of sleepiness on daily life.²⁸ A “normal” score has been proposed to be higher than 18.¹⁷

The Calgary Sleep Apnea Quality of Life Index (SAQLI) is an interviewer-administered questionnaire evaluating the effect of sleep apnea on daily activities, emotional functioning, social interactions, and symptoms over the past 4 w. If subjects are receiving treatment for OSA, an additional subscale includes treatment-related symptoms. Subjects' responses range from 1, which indicates a large effect (e.g. “all-of-the-time”), to 7, which indicates no effect (e.g. “none-of-the-time”) for 56 items. The four subscales results are averaged (1-7) and the total score is an average of the subscales (1-7). A greater SAQLI score indicates less of an effect of OSA on quality of life.

Analysis

To assess the internal reliability of the FOSQ and SAQLI in the HomePAP population, we calculated the Cronbach α coefficient for each of the five subscales of the FOSQ¹⁰ and for the four SAQLI subscales as well as for the overall FOSQ and SAQLI total scores.¹² Data from the baseline examination were used for internal reliability assessment, and included all subjects enrolled (regardless of their subsequent eligibility for continuation in the trial). We also conducted these analyses on the sample eligible for CPAP with data for 3-mo follow-up.

To assess criterion validity, we assessed the correlation between baseline FOSQ and SAQLI subscales and SF-36 domains using the pairwise Pearson correlations because the scales had a relatively normal distribution. Construct validity was assessed by evaluating the correlation of the FOSQ and SAQLI scores with the ESS,²⁶ an extensively used measure of sleep propensity. We also evaluated the correlation of the scales with each other. We assessed for an association of the FOSQ and SAQLI summary scores with AHI severity. AHI categories—defined as 0-14.9, II = 15-23.9, III = 24-49.9 and IV = > 50—were evaluated using the Cochran-Armitage test for trend.

To assess the responsiveness of the FOSQ and SAQLI to CPAP therapy, we calculated the change in the FOSQ and SAQLI subscores and summary scores from baseline to 3-mo follow-up, as well as computed the effect-size, the standardized response mean, and performed paired t-tests. We evaluated for differences in responsiveness of the FOSQ and SAQLI summary scores by CPAP adherence (> 4 h versus less), sex, race (white versus nonwhite), AHI quartile, education (more than high school versus high school or less), ZIP code SES (lowest quartile versus others), self-reported short sleepers (< 6 h versus others) and self-reported long sleep latency (> 30 min versus other) and depressive symptoms (CES-D > 16 versus less)²⁹ using two sample t-test and one-way analysis of variance. Residential SES was derived from multiple census SES measures of the subject's ZIP code as described previously.³⁰ We developed predictive models of FOSQ and SAQLI responsiveness by performing stepwise linear regression. We considered the following as possible explanatory variables: CPAP adherence (> 4 h), demographics, baseline sleepiness (ESS), body mass index (BMI), AHI quartile, depressive symptoms, sleep latency and duration, randomization assignment, and FOSQ and SAQLI total scores at baseline for inclusion in the model.

Finally, we used distribution-based methods to determine the MID in FOSQ and SAQLI scores because the study did not include a suitable anchor question. The standard error of measurement (SEM) reflects the error in the observed score compared to true score and is sample-independent. We used one SEM to represent a meaningful intraindividual HRQL instrument change.³¹ We calculated the SEM using baseline values for the standard deviation (SD) and reliability among those who followed up at 3 mo with the formula: SEM = SD × √(1 − α).

RESULTS

Sample

Baseline enrollment FOSQ, SAQLI and SF-36 questionnaires data were available for 335 subjects. At enrollment, the subjects were a mean age 46.6 y (SD 12.3), 61% male, with BMI mean 37 (SD 8.6), 41% had a college degree, and 66% were white; 21% were black (Table 1). Of the 335 subjects who completed their screening diagnostic study, 191 qualified for continuation in the study with an AHI ≥ 15 on their diagnostic polysomnography (PSG). One hundred thirty-five subjects (71%) returned for follow-up at 3 mo, completed subsequent FOSQ and SAQLI questionnaires and had CPAP adherence data. Compared to the characteristics from the original screening sample, subjects available at 3-mo follow-up were older, more obese, and had shorter sleep time, (P < 0.05 for differences) but did not differ statistically by baseline ESS, total FOSQ, or SAQLI scores. The SF-36 vitality scores were the lowest at baseline and improved the most substantially at 3 mo by a mean of 13.4 (95% confidence interval [CI] 11.6, 15.2); all domains improved over 3 mo (Table 2). Questionnaire completion rates were similar for all instruments with the exception of the FOSQ sexual intimacy items (missing n = 57). The FOSQ and SAQLI subscores were all significantly higher at 3 mo (Table 3).

Table 1.

HomePAP study sample baseline characteristics.

Table 2.

Baseline and 3-mo Medical Outcomes Study 36-Item Short Form Survey subscales scores in HomePAP trial participants, reported as mean (standard deviation).

Table 3.

Functional Outcomes of Sleep Questionnaire and Sleep Apnea Quality of Life Index scores, mean (standard deviation) at baseline and 3 mo follow-up, reliability and calculated standard error of measurement or minimally important difference among subjects with 3 month follow-up data (n = 135).

Internal Reliability

When including all enrollees, the Cronbach α coefficient was 0.94 for the 30-item FOSQ total score. By subscales, the reliabilities were: activity level 0.90 (nine items), vigilance 0.86 (seven items), intimacy 0.85 (four items), general productivity 0.87 (eight items), and social outcome 0.82 (two items). The SAQLI total score (56 items) had a reliability of 0.92. The Cronbach α for the SAQLI subscales were: daily functioning 0.89 (11 items), social interactions 0.88 (13 items), emotional interaction 0.91 (11 items) and symptoms 0.70 (21 items). The reliability was slightly higher when baseline responses were restricted to those with 3-mo follow-up results (Table 3 reliability).

Criterion Validity

Baseline FOSQ and SAQLI sub-scores were significantly correlated with all SF-36 domains, (P < 0.001; Table 4). The FOSQ activity level was moderately correlated with all SF-36 domains except bodily pain. The SAQLI emotional functioning subscale and SF-36 mental health domain were strongly correlated (r = 0.65). The SF-36 social functioning and vitality domains were moderately correlated with all SAQLI subscales (r = 0.40-0.58). The FOSQ and SAQLI subscores and summary scores were also significantly correlated with each other, especially the SAQL daily functioning and FOSQ activity level sub-scores (r = 0.65) (Table 5).

Table 4.

Pearson correlation coefficients for baseline Medical Outcomes Study 36-Item Short Form Survey domains with baseline FOSQ and SAQLI subscales in the HomePAP population (n = 333); all P < 0.001.

Table 5.

Correlation of the FOSQ and SAQLI Subscales and summary scores (n = 333); all P < 0.001.

Construct Validity

The ESS was inversely moderately correlated with the FOSQ vigilance score (r = -0.56, P < 0.001; Table 6) as well as the FOSQ total score (r = -0.42; P < 0.001). The SAQLI subscores were all significantly correlated, albeit generally only weakly, with the ESS (Table 6).

Table 6.

Baseline Epworth Sleepiness Scale Pearson correlation coefficients in the HomePAP population (n = 334), all P < 0.001; with Functional Outcomes of Sleep Questionnaire subscales and total score and Sleep Apnea Quality of Life Index subscales and total score.

Relation with AHI

The baseline FOSQ and SAQLI summary and subscale scores did not differ by AHI quartile, with the exception of the SAQLI emotional functioning subscore, which had a positive test for trend (P = 0.005) with higher scores in the higher AHI quartile. The AHI did not correlate with either the FOSQ or SAQLI summary or subscale scores, except weakly with the SAQLI emotional function (r = 0.134, P = 0.02).

Responsiveness

After 3 mo of CPAP therapy, individual FOSQ total scores increased by a mean of 3.3 points (95% CI: 2.86, 3.81), P < 0.001. The effect size was large at 1.15; the calculated SEM was 0.75 (Table 3). The FOSQ vigilance and activity level subscores had a greater responsiveness than the other subscores (Table 7). After 3 mo of CPAP, the SAQLI total score increased by a mean 0.82, (95% CI 0.65, 0.99), P < 0.001. The effect size was also large at 1.0; the SAQLI total score SEM was 0.29. The daily functioning and social interaction subscores had greater responsiveness than the others (Table 7). Using the 1 SEM cutoff, 86.6% of subjects were above the MID threshold for improvement in the FOSQ total score and 74.3% the SAQLI total score (Figure 1).

Table 7.

Responsiveness of Functional Outcomes of Sleep Questionnaire and Sleep Apnea Quality of Life Index subscores to 3 mo of continuous positive airway pressure therapy in HomePAP population, n = 135.

Figure 1.

Plot of the cumulative distribution of Functional Outcomes of Sleep Questionnaire (FOSQ) and Sleep Apnea Quality of Life Index (SAQLI) change scores. The symbols mark the net difference in total score from baseline to 3 mo. SEM, standard error of measurement.

The FOSQ total score change was significantly higher in those CPAP adherent (> 4 h/day) but the SAQLI total score change was not. Those with a mean of 4 h or more of CPAP use per night (n = 63) increased their FOSQ total score by a mean of 3.9 (SD 2.8) compared to those averaging less than 4 h (n = 72), who had a mean increase of 2.9 (SD 2.8), P = 0.039, (Table 7). In bivariate analysis, the FOSQ total score responsiveness was also significantly greater in women, whites (versus nonwhites) and those with baseline depressive symptoms. The SAQLI total score change was greater in those with baseline depressive symptoms but did not differ by demographics. There was no difference in responsiveness by AHI quartile or sleep habits.

In stepwise regression models, CPAP use, baseline depressive symptoms, and sleepiness were predictive of FOSQ responsiveness (Tables 8 and 9). A baseline CES-D > 16 predicted of an increase of 3.1 (SD 0.6) in FOSQ and of 0.54 (SD 0.24) in SAQLI total score. Demographic variables, baseline sleep habits, BMI, and AHI quartile were not predictive (P > 0.10) of FOSQ or SAQLI responsiveness. If baseline HRQL was included in the models, depressive symptoms and ESS were no longer predictors.

Table 8.

Predictors of Functional Outcomes of Sleep Questionnaire total score change after 3 mo of continuous positive airway pressure use, stepwise multivariate regression, n = 134, R² = 0.51.

Table 9.

Predictors of Sleep Apnea Quality of Life Index total score change after 3 months of continuous positive airway pressure use, stepwise multivariate regression, n = 133, R² = 0.25.

DISCUSSION

These intervention data from a clinical trial provide further evidence that the FOSQ and SAQLI are responsive to CPAP treatment for 3 mo. In a detailed assessment of several psycho-metric properties, we found that both the FOSQ and SAQLI demonstrated excellent internal reliability and had appropriate correlation with the well-established SF-36. The instruments demonstrated strong criterion validity as evidenced by moderate correlations with SF-36 social role and physical role domains. The social domains in SF-36, FOSQ, and SAQLI appear to measure similar constructs of social well-being. Surprisingly, the FOSQ vigilance was only weakly associated with the SF-36 vitality. The SAQLI showed much stronger correlations with the emotional role and mental health domains of the SF-36 than the FOSQ.

The FOSQ and SAQLI each were correlated with the ESS, albeit weakly. Other studies, one of community-dwelling elderly and another of sleep clinic patients, have found more robust association with the FOSQ and ESS, but with a similar stronger association of the vigilance scale and ESS.^28,32 Our correlations were likely impacted by our study cohort inclusion criterion of ESS > 12, constraining the distribution. Daytime sleepiness is a key subjective measure of OSA severity. The FOSQ, which was specifically developed to measure sleepiness related function, correlated with the ESS more than the SAQLI, which was designed to measure a broader impact of OSA on HRQL. Of interest, the FOSQ and SAQLI instrument scores did not differ by AHI quartile. A lack of association of the FOSQ and AHI is consistent with prior studies, and underscores the frequent weak association between objective measures of physiological disturbances and the subjective patient reported outcomes.^8,9,33

The FOSQ demonstrated a slightly greater responsiveness to CPAP therapy. The FOSQ total score, activity, and vigilance subscores effect sizes (responsiveness to CPAP) were large and more substantial than in prior studies.^4,14 Over 4 h of CPAP use predicted greater improvement in the FOSQ total score. A dose response of greater FOSQ improvement with greater CPAP use was seen, as previously demonstrated by Weaver et al.²² The mean FOSQ total score improvement of 1 unit (SD 0.43) associated with CPAP use is above the SEM or proposed MID. As the 4 h per night metric is typically used to define minimal adherence for therapeutic benefit, a 1-unit difference in the FOSQ may be significant clinically. Although the SAQLI did show responsiveness to CPAP therapy, the change was not associated with the amount of CPAP use. Some of the observed HRQL responsiveness may represent a regression to the mean and not the use of CPAP per se.

The interpretation of the FOSQ and SAQLI responsiveness is not well defined; it remains undetermined what magnitude of change is clinically significant. We propose that an intra-individual change in the FOSQ and SAQLI of 1 SEM reflects a minimally important difference of clinical significance. The MID reflects the smallest score difference in a patient-reported outcome that is perceived by the patient as a change (harmful or beneficial).³⁴ Most subjects met these criteria, many reaching twofold to threefold above the SEM. Further psychometric evaluation is needed to guide the application and interpretation of these instruments in clinical trials and clinical practice. As the MID depends not only on the instrument but also the sample and contextual characteristics, using samples from subjects enrolled research trials with similar sleep characteristics as well as more diverse, “real world” clinic patients will provide a possible MID range and error estimate. Additional studies using anchor-based methods³⁵ and other distribution approaches in different populations³⁴ will be useful to confirm our proposed MID based on the SEM.

The FOSQ and SAQLI instruments' responsiveness to CPAP therapy did not differ by baseline demographics, self-reported sleep habits, or AHI. Baseline depressive symptoms were strong predictors of HRQL responsiveness and strongly correlated with baseline HRQL. Thus, the independent effect of depression cannot be evaluated. Responsiveness of these HRQL instruments does not appear to be affected by personal factors such as BMI, AHI, race, education, or residential SES.

Our study population and analysis have several limitations. Subjects all had a minimal ESS of 12, limiting the ability to assess the correlation in individuals with a full range of sleepiness scores, and limiting the ability to evaluate whether non-sleepy OSA subjects have distinct HRQL responses to treatment. The relatively modest sample size at 3 mo follow-up (135 subjects with CPAP data) may lead to a sampling biases. Because all subjects received CPAP therapy (no control arm), we cannot conclude the response was solely an effect of CPAP. The change may also reflect primarily depressive symptom improvement on HRQL ratings; the CES-D was very responsive to CPAP therapy. There is also a potential of residual AHI and other unmeasured confounders contributing to HRQL score changes. Although we conducted multiple comparisons, which increase the risk for a type I error, the overall consistency of the reported associations mitigates a concern about inappropriate inferences. The potential generalizability of the findings across the AHI range is supported by use of data from the enrolled subjects who both did not have moderate OSA as well as those with more severe OSA. The reliability and validity results were similar when only eligible subjects were included in the analysis (data not shown/Table 3).

In summary, our results indicate that in patients with OSA and a minimal ESS score of 12, both the FOSQ and SAQLI demonstrate good internal reliability, criterion validity with moderate correlation with the SF-36, and responsiveness to intervention. These instruments provide information distinct from the AHI. Furthermore, responsiveness of instruments did not differ by demographic factors or sleep habits supporting their use across diverse populations. Although the FOSQ demonstrated a slightly greater responsiveness to CPAP than the SAQLI, this may be because of the selection of patients with an elevated ESS on enrollment. Future research on OSA subjects with and without sleepiness is needed to confirm these findings. Nonetheless, our psychometric analyses support the utility and validity of both instruments for assessing key patient reported outcomes in OSA.

DISCLOSURE STATEMENT

This was not an industry supported study. Support for this study was provided by the American Sleep Medicine Foundation 38-PM-07 Grant: Portable Monitoring for the Diagnosis and Management of OSA. Dr. Rosen has consulted for for Jazz Pharmaceutical. Dr. Auckley research support from Teva and ResMed. Dr. Benca has consulted for Sanofi-Aventis and Merck. Dr. Iber has provided uncompensated consultation services to Apnex Medical. Dr. Foldvary-Schaefer has received research support from CleveMed and UCB, Pharma and has participated in speaking engagements for Jazz Pharmaceuticals. Dr. Zee has consulted for Sanofi-Aventis, Merck, Philips-Respironics and Purdue. Dr. Redline has received research support from Philips-Respironics and ResMed Foundation. The other authors have indicated no financial conflicts of interest.

ABBREVIATIONS

AHI

apnea-hypopnea index

CES-D

Center for Epidemiologic Studies Depression Scale

CPAP

continuous positive airway pressure

ESS

Epworth Sleepiness Scale

FOSQ

Functional Outcomes of Sleep Questionnaire

HRQL

Health Related Quality of Life

OSA

obstructive sleep apnea

MID

Minimally Important Difference

RCT

randomized controlled trial

SAQLI

Sleep Apnea Quality of Life Index

SEM

standard error of measurement

SF-36

Medical Outcomes Study 36-Item Short Form Survey