Cues to Starting CPAP in Obstructive Sleep Apnea: Development and Validation of the Cues to CPAP Use Questionnaire

Abstract

Background:

The reasons that a patient has to start treatment, their “Cues to Action,” are important for determining subsequent health behaviors. Cues to action are an explicit component of the Health Belief Model of continuous positive airway pressure (CPAP) acceptance. At present, there is no scale available to measure this construct for individuals with obstructive sleep apnea (OSA). This paper aims to develop, validate, and describe responding patterns within a sample of patients with OSA to the Cues to CPAP Use Questionnaire (CCUQ).

Method:

Participants were 63 adult patients diagnosed with OSA who had never tried CPAP when initially recruited. The CCUQ was completed at 1 month after being prescribed CPAP.

Results:

Exploratory factor analysis (EFA) showed a 3-factor structure of the 9-item CCUQ, with “Health Cues,” “Partner Cues,” and “Health Professional Cues” subscales accounting for 59.91% of the total variance. The CCUQ demonstrated modest internal consistency and split-half reliability. The questionnaire is brief and user friendly, with readability at a seventh-grade level. The most frequently endorsed cues for starting CPAP were Health Professional Cues (prompting by the sleep physician) and Health Cues such as tiredness and concern about health outcomes.

Conclusions:

This study validates a measure of an important motivational component of the Health Belief Model. Health Professional Cues and internal Health Cues were reported to be the most important prompts to commence CPAP by this patient sample.

Citation:

Olsen S; Smith S; Oei TPS; Douglas J. Cues to starting CPAP in obstructive sleep apnea: development and validation of the cues to CPAP use questionnaire. J Clin Sleep Med 2010;6(3):229-237.

Obstructive Sleep Apnea (OSA) is a common sleep disorder^1,2 for which continuous positive airway pressure (CPAP) therapy is the standard treatment for most patients with moderate to severe disease. Although this treatment is effective in reducing airway obstruction to nonclinical levels, CPAP acceptance (commencement) and adherence is suboptimal.^3,4 The prompts for starting treatment in OSA sufferers include subjective severity of symptoms, but CPAP commencement can also be prompted by the concerns of others, such as a doctor or partner. The aim of this paper is to investigate patients' perceptions of these cues to action in commencing CPAP treatment. Our Health Belief Model (HBM)⁵ research demonstrates that patients begin to develop expectations and beliefs regarding OSA and CPAP treatment early in the treatment process, which then has an impact on treatment adherence.⁶ Patients may have different reasons for starting treatment, and these reasons may impact on their acceptance of CPAP and, perhaps, their eventual treatment adherence in different ways.

BRIEF SUMMARY

Current Knowledge/Study Rationale: It has been demonstrated that specific health beliefs impact on CPAP treatment adherence, especially when measured in the context of a theoretical model. This study validates a measure of “Cues to Action,” which is an explicit component of the Health Belief Model of CPAP acceptance.

Study Impact: The Cues to CPAP Use Questionnaire (CCUQ) is a valid measure of patient cues to start treatment. The CCUQ may be used in the assessment of the Health Belief Model in CPAP commencement and adherence research.

The HBM is predictive of engagement in preventative health behaviors (such as wearing a bicycle helmet and attending health-screening programs)^7,8 but has also been extensively investigated in the prediction of adherence in several disease models^7,9–12 and has been specifically applied to the problem of CPAP adherence.^6,13,14 Figure 1 presents our theorized HBM for CPAP adherence (see Olsen et al.⁵ for a full description of the model).

Figure 1.

A conceptual model of continuous positive airway pressure (CPAP) acceptance and adherence^5,6

RDI refers to respiratory disturbance index; BMI, body mass index, ESS, Epworth Sleepiness Scale.

According to the HBM, engagement in a health behavior is based on readiness to act and the expected benefit of the treatment.^5,6 Readiness to act is determined by perceived susceptibility to the illness or illness consequences if the problem is left untreated, as well as the perceived severity of the illness.^8–10,12 Susceptibility to the illness may be articulated in the context of OSA in the patients' belief in the accuracy of the diagnosis and their belief in the risk of experiencing poor health as a result of untreated OSA. Perceived severity of the condition is based around the individual's perception of the impact the illness has on important areas of health and lifestyle functioning. In OSA, these effects may be evident in various health comorbidities but, perhaps more importantly, the effect of sleepiness on proximally important activities of daily life. That is, the severity construct reflects the effect of OSA on work, family life, general productivity, and social relationships.⁷

In this context, the patients with sufficient perception of risk and severity of their OSA will engage in the recommended health behavior (CPAP use) if they have sufficient belief in the effectiveness of the treatment option presented in reducing these potential threats to their health and lifestyle. That is, the individual perceives that there will be sufficient positive consequences (benefits) of using the treatment. Barriers to action are the patient's perception of the potential or actual negative aspects of the treatment, such as side effects, social stigma, inconvenience, cost of the treatment or other difficulties with the treatment.⁷ The ratio of barriers to benefits is important in determining health action. According to the HBM, at least 1 or more cues to action are required to act as a “trigger” to commence the health behavior.⁸ Self-efficacy is theorized to be an important construct in explaining more complex adherence behaviors. Self-efficacy (confidence) is defined in terms of being able to successfully adopt and use the treatment in the face of barriers to its use.^8,10

It is theorized that the HBM is useful in determining initial CPAP acceptance. Once patients have commenced on the treatment, their ongoing experience flows back through the health belief constructs to determine future adherence. As such, the HBM may be assessed at various time points, from before CPAP acceptance to several months to years after commencement (adherence) (see Olsen et al.⁵). At present, validated measures have been identified for all the constructs of the model,⁶ with the exception of the barriers and cues to action constructs. Our research group has developed measures for both of these constructs, with the current study devoted to describing the development and validation of the cues-to-action construct. We are currently also testing the ability of the HBM to predict CPAP commencement and adherence across time.

Despite being an explicit and critical feature of the model, very few HBM studies across the gamut of health behavior research have developed and assessed measures of the cues-to-action construct.^7,13,15–18 This is likely due to the fact that cues to action is a difficult construct to assess prospectively (that is, before engagement in the behavior). Thus, it is often measured on a cross-sectional basis as a means of retrospectively understanding “prompts” to seek help or to commence the health behavior in individuals who have already undertaken the required behavior.^15,18–20 Theoretically, in combination with the ideal balance of HBM constructs of barriers, benefits, severity, risk, and self-efficacy to promote readiness to accept the treatment, a cue must then trigger the actual commencement of the required behavior. Thus, cues to action may be considered after exposure to the treatment as a potential predictor of its use. Cues to action may be internal (such as noticing difficult symptoms), external (for example, mass media campaigns, advice from health practitioner), or interpersonal (such as encouragement from the spouse, advice from friends or family).^7,8

Measures of the cues-to-action construct within the HBM have not previously been validated, and it has therefore not been possible to fully test the HBM in OSA. Four relevant studies have investigated patient-reported cues to action in commencing on CPAP.^13,21–23 Important factors likely to be cues to action may be the desire to overcome health-related concerns, such as abnormal fatigue and somnolence.²¹ A recent study of patients with OSA in Israel found that patients who went on to accept CPAP treatment reported that they had received positive infor-mation about CPAP treatment from family and friends.²² Sage and colleagues¹³ investigated the HBM in OSA adherence, utilizing a measure of cues to action that included perceived health cues, partner prompts, referral by a physician, or referral by a friend. They found that cues to action did not aid in adherence prediction. However, the measures used in their study were not empirically validated. Some work has also been done in the area of the source of patient referral for OSA treatment. Hoy and colleagues²³ found that patients classified as “self-referrers” were more likely to have a greater Epworth Sleepiness Scale (ESS) score than patients who were “partner referred” and that self-referrers were more likely to adhere to the treatment. Approximately 50% of the sample was self-referred, whereas the remaining 50% were referred on the basis of their partner's prompting.

Outside of the OSA literature, cues to action have been investigated in areas such as prostate cancer screening,¹⁸ mammography screening,¹⁹ pelvic floor exercise adherence,²⁰ and general cues to health-promoting behaviors.¹⁵ Important cues to action in these disease models include physician recommendation and other health-care providers as sources of risk-related information, having a family member or friend with the disease, advice from family members or friends, public announcements through radio and television,^15,18 and a perception of the need to change one's health status.¹⁵ It is likely that cues to action are just as important for patients with OSA. Based on the available literature, the likely domains of interest in the measurement CPAP commencement cues may include cues related to health concerns (classed as “internal cues”), cues associated with health professional recommendations, cues from the partner, and recommendations from family and friends.

To integrate previous findings on CPAP use in OSA, and to fully test the HBM in OSA, the Cues to CPAP Use Questionnaire (CCUQ) was developed. The validation and pattern of responding among patients who were newly diagnosed with OSA and recommended for CPAP treatment are presented.

METHOD

Participants

Participants were 63 patients from the larger sample of 113 adult patients diagnosed with OSA^1,24 recruited through the Sleep Disorders Centre at The Prince Charles Hospital, in Brisbane, Australia. These patients responded to a questionnaire mail out 1 month after being prescribed CPAP treatment (56% response rate). Patients who responded to the mail out were a little older (58.51 vs 50.00 years of age, p < 0.001) and scored lower on the ESS at pretreatment (10.35 vs 12.16, p = 0.04) than nonresponders. There were no other demographic or biomedical differences between responders and nonresponders. All participants followed the same initial time course and intervention (standard care at The Prince Charles Hospital).

All patients met International Classification of Sleep Disorders-II² criteria for the diagnosis of OSA. Inclusion criteria were having diagnosis of OSA, with a clinical recommendation for CPAP treatment, but having never tried CPAP treatment. Patients were excluded from the study if they were less than 18 years of age, were found to require bilevel ventilation (e.g., due to evidence of central sleep apnea syndrome), did not complete CPAP titration, or were unable to give informed consent. This study had ethics approval from Human Research Ethics Committees of the University of Queensland and the Prince Charles Hospital Health Services District. All participants gave informed consent to participate.

Table 1 displays the demographic characteristics of the sample, which was predominantly male (37 male patients, 59% of sample), married, middle-aged, obese, and, on average, diagnosed with severe OSA (respiratory disturbance index [RDI] a measure of the total number of partial (hypopneas) or complete (apneas) disturbances in breathing per hour of sleep > 30 events/h).²⁴ The sample is similar demographically to those of previous studies,^3,4,6,25,26 suggesting that this sample is fairly “typical” of the OSA population.

Table 1.

Demographic characteristics of the patient sample

	Mean (SD)	Range
Age, y	58.5 (10.8)	27.0–79.0
RDI, events/h	36.7 (22.8)	5.0–94.8
BMI, kg/m2	36.4 (8.4)	22.0–67.0
ESS score	10.4 (4.9)	0.0–21.0
Years of education	11.4 (3.5)	4.0–22.0
Relationship statusa
Married	60
De facto relationship	5
Relationship not living together	0
Single	35

^aData are presented as mean percentages.

RDI refers to respiratory disturbance index; BMI, body mass index; ESS, Epworth Sleepiness Scale.

Materials

Instrument Development

Items for the CCUQ were developed from review of treatment-seeking literature and with consultation with an expert panel. All members of the panel had direct experience with the OSA population and CPAP therapy. The panel consisted of 3 sleep physicians (author JD with 7 years experience in sleep medicine at the time of item generation, a consultant with 3 years of specific experience, and a sleep fellow with 1 year of ongoing experience), 2 clinical nurse consultants (15 and 5 years of experience, respectively), 2 senior medical scientists (with 8 years and more than 20 years of experience, respectively), and 3 specialists in sleep and clinical psychology (authors SS, TO, and SO, who had 15, 2 and 1 year of experience as in sleep psychology, respectively, at the time of item generation). Cues, such as advice from significant others—including friends and family, advice or concern from the spouse or partner, self-identified health concerns, and advice from medical practitioners, were identified in the general treatment-seeking literature^15–18 and CPAP-specific literature.^13,21–23

Twelve pilot items addressing these cue areas were initially developed. After review by the expert panel, a final set of 10 items was chosen (4 items related to personal health concerns, 2 items related to partner/spouse concern/advice, 2 items related to advice from a family member or friend (not spouse/partner), and 2 items related to physician/doctor advice). One item related to mass media or public service announcements was removed on the basis that it is likely to be related to seeking help for OSA, rather than a trigger for the decision to commence on CPAP. One item “I didn't think I had a sleep problem, but a health professional recommended that I seek help” was also removed for the reason that it did not directly target CPAP commencement.

The 10 chosen items were rated on a 4-point Likert scale (0 = not at all, 1 = a little important, 2 = moderately important, 3 = extremely important) measuring the importance of certain cues in the decision to commence CPAP treatment. Total scores are derived from summed item scores. Higher scores indicate higher rated importance of cues to start treatment. An additional space was available on the scale to allow patients to note any additional cues to action that were relevant to them. Nine patients made an additional comment. These included 3 comments related to health concerns (tiredness, problems sleeping), 4 comments related to prompting to commence CPAP by a health professional, 1 comment related to concern of the patient's husband who had witnessed her apneas, and 1 comment related to workplace requirements. Because these additional comments were infrequent in the broader sample and they appeared to fall under the broad domains assessed by the items already in the scale, the measure was not modified.

Additional Assessment

When initially recruited, participants completed a questionnaire battery that consisted of demographic questions, including age, marital status, and whether they had previously used CPAP. This information was verified against medical records.

The baseline ESS²⁷ score obtained at the initial consultation with the sleep physician was used to assess criterion validity. The ESS consists of 8 items are rated on a scale from 0 to 3 (0 = would never doze, 3 = high chance of dozing). Higher scores indicate greater tendency to fall asleep in a variety of contexts. It has high internal consistency²⁷ when assessed with patients with OSA and demonstrates good correlations with objective measures of sleep latency.²⁸ It was expected that higher baseline levels of sleepiness at pretreatment would be associated with greater internal health cues to action²³ but not interpersonal or health professional cues to action.

At the 1-month time point, in addition to the CCUQ, patients also completed 2 measures containing 4 constructs from the HBM, which are available in the literature. These were measures of perceived self-efficacy, risk, severity, and benefits. The psychometric properties of each of these measures are described below

The Self-Efficacy Measure for Sleep Apnea

The Self-Efficacy Measure for Sleep Apnea²⁹ is a 26-item questionnaire consisting of 3 subscales measuring adherence-related cognitions, namely, self-efficacy, risk perception, and benefits (outcome expectancy) (see Figure 1). Items are rated on a Likert scale from 1 to 4, with higher scores indicating greater perceived self-efficacy, greater risk perception, and higher benefits expectancy with treatment. Internal consistencies range from 0.85 to 0.89, and factor analysis confirms the 3 independent subscales.

Functional Outcomes of Sleep Questionnaire

The Functional Outcomes of Sleep Questionnaire³⁰ is a 30-item survey of sleep-related quality of life in OSA and is a measure of perceived severity (see Figure 1). It directly assesses the perception of impact of the disorder on daily functioning. Functional limitations of this nature provide a better severity measure than do other objective measures (such as RDI) because it directly taps patient perceptions of the degree of impact of the disorder on important areas of functioning. Items assessing 5 domains, including activity level, vigilance, intimacy, general productivity, and social outcomes, are rated on a Likert scale ranging from 1 to 4 (1 = yes, extreme difficulty, 4 = no difficulty). Lower scores represent greater impairment in functioning. A mean-centered total score representing total functional difficulties related to sleepiness is calculated. Internal consistency estimates range from 0.81 to 0.90 for the subscales, and is 0.95 for the total score.

There were no other adequate and standardized measures of the HBM cues-to-action construct for the OSA population with which to perform concurrent validity analyses.

Procedure

Patients who met the inclusion criteria for the study were invited to participate by their sleep physician during the follow-up appointment after the patient's diagnostic polysomnogram. The polysomnogram was scored by trained sleep scientists using recommended guidelines.²⁴ The diagnosis of OSA was explained to the patient by their sleep physician, and the recommended treatment option (CPAP) was described. There was no set script for provision of this information, although fairly specific information is covered in the structured 10-minute consultation. This includes review of the results of the diagnostic sleep study using a visual printout of the diagrammatic summary from the diagnostic sleep study. Diagnostic criteria are reviewed and discussed with patients in terms of the predefined levels of OSA severity. Potential consequences of OSA are explained to the patient. Treatment options are presented, including lifestyle modification, posture, mandibular splints, or CPAP, depending on severity of OSA, patients' wishes, and indications and contraindications for certain therapies. CPAP treatment was indicated for every patient in this sample.

Approximately 1 month after the patients received their CPAP prescription, the questionnaire containing the CCUQ was mailed to the patient's home address and included a stamped addressed envelope. The 1-month time point was chosen to coincide with the return of a 1-month standard form that patients were given by the nurse when they received their CPAP prescription. This form included questions about any difficulties with using CPAP, any difficulties with purchasing CPAP, and a request for the patients to contact the clinic if they were experiencing any difficulties. Patients were requested to return the form to the nurse at 1 month using the reply-paid envelope. Use of this time point also allowed the opportunity for most patients to have filled their prescription and to have started on CPAP and, thus, be able to report on cues that did or may prompt their decision to start.

Data Analysis

All data analyses were conducted using SPSS for Windows version 15.0 (SPSS, Chicago, IL). Relationships between variables and item subscales were assessed using Pearson product moment correlation coefficients. Four-level independent groups Functional analysis of variance was used to assess the relationship between the CCUQ subscales and relationship status (posthoc analyses were conducted with Bonferroni correction when significant global differences between groups were identified). Independent group t tests and χ² tests were used to assess potential differences between sample demographic and biomedical variables.

Construct validity of the questionnaires was assessed using exploratory factor analysis (with principal components extraction method). Varimax rotation was requested when more than 2 factors were present. Assessment of the factor solution was based on identification of parsimonious and coherent constructs demonstrating eigenvalues greater than 1.0 and with item-factor loadings greater than 0.40. In addition, items were to be excluded if they loaded significantly (> 0.40) on more than 1 factor. Internal consistency was assessed using Cronbach α reliability coefficients. Split-half reliability was determined using Spearman-Brown split-half reliability coefficients. Higher-order relationships between questionnaire items were checked for violation of assumptions of normality and linearity. Tolerance levels did not fall below acceptable levels, indicating that collinearity and singularity of variables was not present.³¹ Of the 63 participants who completed the follow-up questionnaire, 3 participants missed 1 or more items of the CCUQ. These cases were deleted listwise from the factor analyses, leaving 60 cases for analysis.

RESULTS

The 10 items initially investigated for inclusion in the CCUQ were assessed to check for violations of assumptions of normality and linearity. Two items were found to be moderately skewed. One item (related to encouragement by a family member (not spouse/partner) demonstrated violation of bivariate and multivariate assumptions and so was removed from further analyses. One item (related use of CPAP following advice from a friend/acquaintance who does not have sleep apnea) demonstrated moderate skew, with 80.5%, indicating that this was “not at all” a cue for them. However, this item adhered to assumptions of bivariate and multivariate normality and was retained for further analysis.

Inspection of the scale properties and frequency of responding for these items indicated that none of the remaining items demonstrated a response pattern in which more than 85% of respondents selected any 1 response.

Construct Validity: Exploratory Factor Analysis

Exploratory factor analysis of the remaining 9 items revealed that 59.91% of the total variance was explained by a 3-factor solution (with a likely 3- factor solution supported using criteria of eigenvalues > 1.0 and by inspection of the scree plot). As demonstrated in Table 2, 5 items loaded on factor 1, with good factor loadings well above 0.40. Two items loaded on factor 2, and 2 items loaded on factor 3. Factor 1 “Health Cues” was associated with items such as concern about tiredness, heart complaints, concern about health consequences such as car crashes, and advice from a friend/acquaintance (who does not have OSA). Factor 2 “Partner Cues” was associated with items related to partner encouragement and partner difficulty with sleep due to the patient's snoring. Factor 3, “Health Professional Cues,” consisted of items related to physician concern and advice to use CPAP.

Table 2.

Factor loadings for the CCUQ

Item descriptor	Factor 1: Health Cues	Factor 2: Partner Cues	Factor 3: Health Profess. Cues	Communalities
I was worried about the health consequences of my sleep problem	0.68	–0.12	0.17	0.50
I was so tired all of the time	0.62	–0.36	0.02	0.51
I was worried about my heart	0.67	0.27	0.01	0.52
I was worried that I would have a car accident	0.65	0.18	0.19	0.49
Advice from friend/acquaintance (who does not have OSA)	0.58	0.05	0.03	0.34
Partner encouraged me to start CPAP	–0.05	0.82	0.15	0.69
Partner couldn't sleep because of my snoring	0.17	0.85	–0.04	0.75
My sleep physician said that I should	0.11	–0.02	0.89	0.81
My sleep physician was worried about my OSA	0.14	0.12	0.87	0.79
% of variance explained by factor	23.34	18.39	18.18
Total variance explained by factor solution = 59.91

CCUQ refers to Cues to CPAP Use Questionnaire; OSA, obstructive sleep apnea; CPAP, continuous positive airway pressure.

Extraction method: Principal Component Analysis; Rotated factor solution (varimax).

Scale Characteristics

Reading Ease

The Flesch readability of the CCUQ was 66.1³² (scores between 60 and 70 are considered acceptable) and indicated that the questionnaire was suited to a seventh-grade reading level. The final questionnaire is presented in the Appendix.

Internal Consistency and Item to Total Correlations

The Cronbach α for CCUQ was 0.63, and the Spearman-Brown split-half reliability coefficient was 0.67. Table 3 provides a summary of the questionnaire and subscale properties of the CCUQ.

Table 3.

Summary statistics of the CCUQ scale and subscales

Scale	Mean (SD)	Obtained Range (possible range)	Cronbach α	Item-to-total scale correlations
CCUQ- total score	13.94 (4.88)	4 – 26 (0 – 27)	0.63	0.17 - 0.46
Health Cue	6.92 (3.35)	0 – 15 (0 – 15)	0.60	0.32 - 0.51
Partner Cue	2.15 (2.21)	0 – 6 (0 – 6)	0.70	0.55 - 0.55
Health Professional Cue	4.87 (1.57)	1 – 6 (0 – 6)	0.74	0.60 - 0.60

CCUQ refers to Cues to CPAP Use Questionnaire.

The αs for 3 subscales ranged from 0.60 (Health Cue) to 0.74 (Health Professional Cue). In all, the CCUQ provides an “adequate”^33,34 measure of the Partner Cue and Health Professional Cue constructs and a modest measure of the Health Cue construct in terms of reliability. ³⁴

Concurrent Validity

The relationships between the CCUQ subscales and existing HBM constructs are presented in Table 4. As demonstrated in Table 4, CCUQ total score was significantly positively correlated with benefits and self-efficacy but not with severity or risk perception. Higher importance of Health Cues was significantly positively correlated with greater benefits perception and greater self-efficacy but not with severity or risk. Partner Cues was significantly positively correlated with self-efficacy, as was the Health Professional Cues subscale.

Table 4.

Bivariate correlations between HBM constructs and CCUQ subscales

	CCUQ Total Score	Health Cues	Partner Cues	Health Professional Cues
Severity	–0.09	–0.16	0.01	0.02
Risk	0.05	0.02	0.15	–0.11
Benefits	0.49b	0.47b	0.22	0.20
Self-efficacy	0.43b	0.30a	0.26a	0.29a

HBM refers to Health Behavior Measure; CCUQ, Cues to CPAP Use Questionnaire;

^ap < 0.05;

^bp < 0.01

Criterion Validity

Because there is not currently any comparable measure of cues to action in OSA populations, measures of construct validity could not be completed. Criterion validity was assessed by investigating relationships between the CCUQ subscale scores and related demographic variables of ESS, RDI, and relationship status. It was expected that greater perceived sleepiness may be related to a higher rating of the importance of Health Cues to action. RDI may relate to physician prompting and, thus, a higher judgment of Health Professional cues to action. It was expected that patients in a romantic relationship would rate Partner Cues as more important cues to action than would single individuals.

ESS was significantly correlated with the CCUQ total score (r = 0.33, p < 0.01) and Health Cue score (r = 0.28, p = 0.03) but not Partner (r = 0.14, p = 0.27) or Health Professional Cue (r = 0.22, p = 0.09) scores, indicating that a higher rating of the importance of cues to action and, specifically, Health Cues to action, is associated with higher severity of subjective sleepiness. RDI, a measure of objective respiratory disturbance, was not significantly correlated with any of the CCUQ subscales (Health Cue, r = −0.08, p = 0.53; Partner Cue, r = −0.02, p = 0.91; Health Professional Cue, r = −0.13, p = 0.30) or total score (r = −0.11, p = 0.42). There was a significant difference between Partner Cue score for individuals with different relationship status (F_3,59 = 10.02, p < 0.01, pη² = 0.25). Posthoc analyses revealed that single individuals reported lower importance of Partner Cues than did individuals who were married (p < 0.01) or in de facto relationships (p = 0.01). There was also a significant difference in Health Cue score between individuals of different relationship status (F_3,59 = 3.88, p = 0.03, pη² = 0.12). However, with Bonferroni correction, the difference between married and single individuals did not reach significance (p = 0.10). There were no other significant differences between the groups with the remaining CCUQ scales.

Descriptive Item Characteristics

Frequency of Endorsement of Individual Items

The frequency of endorsement of each item of CCUQ is presented here to provide additional descriptive information for interest of treating practitioners (see Table 5). The most highly endorsed cues for starting CPAP were encouragement by the sleep physician (88.7% rated as moderately or extremely important), worry regarding health consequences of OSA (83.9% rated as moderately or extremely important), physician-indicated concern about the patient's OSA (83.6% rated as moderately or extremely important), and tiredness (67.8% rated as moderately or extremely important). The least-endorsed cues for starting CPAP were concern regarding having a car crash (24.2% rated as moderately or extremely important) and advice from a friend/acquaintance (13% rated as moderately or extremely important).

Table 5.

Frequency of endorsement of CCUQ items

Item descriptor	“not at all”	“a little important”	“moderately important”	“extremely important”
My sleep physician said that I should	3.2	8.1	14.5	74.2
My sleep physician was worried about my OSA	8.2	8.2	27.9	55.7
I was worried about the health consequences of my sleep problem	3.2	12.9	22.6	61.3
I was so tired all of the time	16.1	16.1	17.8	50.0
I was worried about my heart	39.3	16.4	16.4	27.9
Partner encouraged me to start using CPAP	56.5	0.0	17.7	25.8
My partner couldn't sleep because of my snoring	50.0	14.5	19.4	16.1
I was worried that I would have a car accident	59.7	16.1	9.7	14.5
Advice from friend/acquaintance (who does not have OSA)	80.5	6.5	6.5	6.5

Data are presented as percentages.

CCUS refers to Cues to CPAP Use Questionnaire; CPAP, continuous positive airway pressure; OSA, obstructive sleep apnea.

DISCUSSION

The aim of this paper was to develop the CCUQ and describe patient response to this questionnaire at 1 month after initiation pf CPAP treatment. In prior work, we demonstrated the utility of the HBM in the prediction of CPAP adherence with existing validated measures of risk, functional outcomes (severity), outcome expectancies (benefits), and self-efficacy.⁶ In the current paper, we developed and validated the CCUQ as a measure of the cues to action construct from the HBM to promote testing of the full HBM in future studies.

The 9-item CCUQ was comprised of 3 theoretically driven subscales (supported by exploratory factor analysis) assessing Health Cues, Partner Cues, and Health Professional Cues. This questionnaire showed adequate scale internal consistency at a level that is deemed appropriate for measures in the early stages of development and validation.³⁴ Given that the CCUQ is a new scale assessing a largely untapped domain within the HBM for individuals with OSA, it is suggested that further validation of this measure in future studies with larger samples sizes is warranted and may lead to improvement and refinement of the scale (thus improving internal-consistency estimates). The CCUQ demonstrated good readability and is suitable for individuals at a seventh-grade reading level, which both promotes the face validity of the questionnaire and reduces the likelihood of response errors arising due to comprehension difficulties. The scales demonstrated sensible relationships with demographic and biomedical variables, such as subjective sleepiness (ESS) and objective relationship status, supporting the criterion validity of the CCUQ. The CCUQ subscales were moderately correlated with other HBM constructs, such as self-efficacy and benefits, but not correlated with risk or severity perceptions. Overall, these findings suggest that the CCUQ measures a domain within the broader HBM, without overlapping to a large degree with existing measures in the model. Thus, the CCUQ is likely to contribute beneficially to the HBM.

Consistent with research about treatment seeking and health action in other health domains,^15,18 Health Professional cues were rated as most important for starting CPAP, followed by Health cues, such as concern about the health consequences of OSA and tiredness. Advice from a friend/acquaintance and concern about having a car crash were rated as less important cues for commencing CPAP, as compared with seeking help for symptoms. These findings suggest that patients rate advice from health professionals (specifically their sleep physician) as very important in their decision to commence on CPAP. More than 80% of patients indicated that sleep physician-prompting to use CPAP and a perception of sleep-physician concern regarding the patient's OSA were moderately to extremely important cues to commence on CPAP. This suggests that clear communication by the health professional and appropriate education regarding the disorder and treatment may be important to patients in supporting treatment uptake.^35–38

The current sample size was modest, and the exploratory factor analysis of this scale was used to suggest the most likely factor structure. Given the early stages of this questionnaire development, exploratory factor analysis is the most appropriate method for assessment of construct validity. The Health Professional cue and Partner cue subscales have 2 items, which may be addressed in future research. The fact that the items within these subscales demonstrated very high factor loadings (> 0.80) indicates the stability and utility of the subscales as unitary construct within the full questionnaire. Additional items may be added to these scales if needed for stability in future studies. Future studies will also need to further validate this measure earlier in the treatment process (i.e., as close as possible to CPAP initiation). The 1-month time point was chosen in this sample because it coincided with a critical time point of patient contact with the Prince Charles Sleep Disorders Centre. This decision was made to minimize interruption to standard patient care. However, the potential for response bias due to the 1-month gap between treatment commencement and completion of the CCUQ may be addressed by cross-validation of the questionnaire in similar samples that have completed the questionnaire earlier in the treatment process.

A limitation of this current study was the use of a relatively small subset of items at the outset of instrument development. This limitation may be addressed in future studies as further development of the CCUQ, and the HBM in general, is undertaken. In contrast, there were several significant strengths of the approach used, which included review of the literature available in this area and review by an expert panel with several years of specialist clinical sleep medicine experience working in this area. Specifically, the questionnaire has good theoretical grounding and was purposely developed to promote a theoretically derived HBM assessment of CPAP commencement.

There was a modest response rate to the follow-up questionnaires containing the CCUQ (56% return rate). Low response rates to follow-up questionnaires is not uncommon in the health domain.³⁹ The responders were not different from nonresponders on any demographic or biomedical variables other than age and subjective sleepiness on the ESS, with older participants who were less sleepy more likely to respond to follow-up. There is the potential that the impact of Health Cues related to sleepiness may have been lessened as a result of responders to the questionnaire being less sleepy, but there was still a significant correlation between the Health Cues scale and baseline ESS score.

There is further validation work that can and should be completed with the CCUQ, specifically, a confirmatory factor analysis with commensurate larger sample size and testing across different varied OSA populations, such as with samples with varied severity of OSA, varied levels of pretreatment symptoms, and location at different sites. A specific research direction for our team is the use of these validated measures in the assessment of the HBM in the early prediction of CPAP adherence.

In summary, we have developed and validated a measure designed to facilitate assessment of the HBM in CPAP commencement and adherence research. The CCUQ is brief user-friendly (comprehensible to individuals with a seventh-grade reading level) and theory driven. Health Professional cues and Health concerns were the most important cues, as reported by patients in prompting them to commence CPAP.

DISCLOSURE STATMENT

This was not an industry supported study. The authors have indicated no financial conflicts of interest.

APPENDIX

Cues to CPAP Use Questionnaire (CCUQ)

In this section we would like you to indicate how important the following factors were in your decision to start using CPAP. If a particular statement is not applicable for you, please indicate this by filling in the “⓪ not at all” response option. The scale is provided below: