Abstract
Infants, children, and adolescents are increasingly being prescribed continuous positive airway pressure (CPAP) for treatment of obstructive sleep apnea syndrome (OSAS), yet adherence is often poor. The purpose of this study was to examine the relationship between caregiver and patient-reported health cognitions about CPAP prior to starting CPAP and CPAP adherence at 1 month. We hypothesized that greater caregiver-reported self-efficacy would be positively associated with CPAP adherence in children. We also evaluated patient-reported self-efficacy and caregiver- and patient-reported risk perception and outcome expectations as they related to adherence, as well as how demographic factors influenced these relationships.
A pediatric modification of the Self-Efficacy Measure for Sleep Apnea Questionnaire was administered to children and adolescents with OSAS-prescribed CPAP and their caregivers during the clinical CPAP-initiation visit. The primary outcome variable for adherence was the average total minutes of CPAP usage across all days from the date that CPAP was initiated to 31 days later.
Unadjusted ordinary least-square regression showed a significant association between caregiver-reported self-efficacy and adherence (p = .007), indicating that mean daily CPAP usage increased by 48.4 minutes when caregiver-reported self-efficacy increased by one point (95% confidence interval 13.4–83.4 minutes). No other caregiver- or patient-reported cognitive health variables were related to CPAP use.
This study indicates that caregiver CPAP-specific self-efficacy is an important factor to consider when starting youth on CPAP therapy for OSAS. Employing strategies to improve caregiver self-efficacy, beginning at CPAP initiation, may promote CPAP adherence.
Obstructive sleep apnea syndrome (OSAS) is increasingly being diagnosed in children. For those in which surgery is not indicated or for those in which residual OSAS persists, CPAP is an efficacious treatment but adherence is a challenge. Identifying modifiable factors that influence adherence is of paramount importance to improve treatment effectiveness. This study suggests that caregiver-reported self-efficacy specific to CPAP, the belief that a caregiver has the ability to successfully implement and use CPAP with his/her child, is an important predictor of short-term CPAP adherence. Caregiver-reported self-efficacy is a modifiable factor that can be addressed by clinicians to improve adherence in their patients.
INTRODUCTION
Obstructive sleep apnea syndrome (OSAS) is a common pediatric disorder with prevalence estimates between 2% and 10%.1 Many youth who are at highest risk for OSAS are impacted by comorbid conditions (eg, craniofacial anomalies, Down syndrome, Prader Willi syndrome, obesity) and will live with OSAS for much of their lives. Their health and development may be particularly impacted by untreated OSAS.
Continuous positive airway pressure (CPAP) therapy is often prescribed for patients with OSAS who have residual OSAS following adenotonsillectomy, who are not candidates for adenotonsillectomy, or whose parents decline the surgery. CPAP therapy is a safe and effective treatment for OSAS.2 While CPAP therapy is highly efficacious, adherence is poor in both adult and pediatric populations.2–4 Approximately 50% of adult patients prescribed CPAP are nonadherent with therapy recommendation at 1 year, and 15%–30% of patients refuse CPAP treatment prior to ever receiving a CPAP machine.5,6 Further, in adults, it has been described that the majority of CPAP dropouts occur early in treatment, with fewer patients discontinuing it as time persists.6 CPAP therapy in children and adolescents may be even more difficult to tolerate, particularly in those with comorbid medical or developmental conditions. Reported CPAP adherence rates in pediatric studies range from 33% to 98% depending on the definition of adherence.2,7,8 Similar to adults,7 usage in the first week of treatment has been found to predict longer term use over 2 to 3 months in children in some7 but not all studies.2 Therefore, identification of predictors, and especially modifiable predictors, of adherence prior to starting CPAP is of paramount importance in improving adherence at the initiation of CPAP and to mitigate physical and neurobehavioral consequences of untreated OSAS in youth.
There are few studies examining factors that impact adherence to CPAP in a pediatric population.4,9 In children, caregiver reported self-efficacy may be an important predictor of adherence as caregivers are often the agents of health behaviors and implementation of pediatric medical interventions. Self-efficacy is defined as belief in one’s ability to successfully perform a specified behavior or set of related behaviors10 and includes both knowledge of and confidence in one’s ability to perform tasks.4,11,12 Investigators have previously utilized the Self-Efficacy Measure for Sleep Apnea (SEMSA) in adults to evaluate self-efficacy and adherence to CPAP therapy.13,14 The aim of this study was to examine the relationship between health cognitions, specifically self-efficacy and short-term adherence to CPAP. The primary hypothesis was that greater caregiver-reported self-efficacy would be positively associated with CPAP adherence at 1 month following the initiation of CPAP. We also evaluated the relationship between risk perception of OSAS and outcome expectancies of CPAP use and short-term adherence to CPAP, as well as how demographic factors influenced those relationships.
METHODS
As part of a quality improvement effort, an interdisciplinary team (ie, physicians, nurses, respiratory therapist [RT], and psychologist) developed and implemented a semi-structured, individually tailored model for initiating CPAP therapy during an outpatient visit in the Sleep Center at a large, urban pediatric hospital. This process has been described in detail.15 The intervention involved education regarding OSAS, causes and consequences of untreated OSAS, and risks and benefits of CPAP therapy; as well as detailed instruction and demonstration of use and care of the equipment. The child was fit for a mask and tried on the CPAP at pressure. Anticipatory guidance was provided for technical issues and potential side effects, and the psychologist provided anticipatory guidance regarding the behavioral implementation of CPAP. Further, the psychologist and family developed an individualized behavioral plan for implementing CPAP in the family’s home. In addition, the psychologist demonstrated behavioral techniques with the child and the CPAP during the visit as a role model for the caregivers. Following the initiation process, the caregiver and the child, if the child was ≥10 years with no reported developmental delays, completed a questionnaire that assessed self-efficacy for CPAP and additionally yielded scores for perception of risk and outcome expectations.
A waiver of consent was approved through the institutional review board, as these are retrospective data obtained from a quality improvement initiative. The institutional review board of the hospital approved analysis of the data.
Self-Efficacy
Self-efficacy was assessed with the SEMSA modified by its creator and pediatric colleagues for caregivers and children. The SEMSA for caregivers is a 20-item questionnaire assessing CPAP adherence-related cognitions based on principles of social cognitive theory.13 The instrument is divided into three subscales that measure risk perception of OSAS (eg, cardiovascular risk), outcome expectancies of CPAP use (eg, improved alertness), and treatment-perceived self-efficacy (the individual’s confidence in using CPAP treatment despite challenges). Items are rated from 0 to 4 on a Likert scale (0 = very low and 4 = very high). The mean of the nonmissing item responses was calculated for each of the three subscales. Higher scores indicate greater risk perception, higher outcome expectancies with treatment, and greater perceived self-efficacy, respectively. The questionnaire takes approximately 15 minutes to complete. The SEMSA for children is similar in questions and format as that of the caregiver but is a self-report version. Caregivers and children completed the questionnaires independently; while they completed the questionnaires in the same room, they were monitored by a study team member to minimize interaction.
Adherence
Data cards from the CPAP device were downloaded or device modems were accessed at follow-up clinic visits using EncoreAnywhere™. The primary outcome variable for adherence was the average total minutes of CPAP usage across all days from the date that CPAP was initiated to 31 days later. Percentage of days used and average minutes of usage on days used were also calculated.
Statistical Analysis
Statistical analysis was performed with R version 3.2.4 (R Foundation for Statistical Computing, www.R-project.org). Data were summarized as median and interquartile range (IQR) for quantitative variables and as frequency (%) for categorical variables. Pairwise correlations within adherence variables and within SEMSA variables were evaluated using Pearson correlations.
The predictive ability of the SEMSA for the mean adherence was evaluated using linear regression. Following the unadjusted regression of adherence on the questionnaire results, potential confounders such as age, sex, race, obesity, and developmental disability were examined by including their main effect terms and interaction terms with the questionnaire data. As a sensitivity analysis to examine the robustness of the ordinary least square regression (OLS) against the skewness in the data, quantile regression was performed for median adherence. In addition, logistic regression was utilized to evaluate the association between nonusers [dichotomized at 20 minutes mean use per night (Figure 1)] and partial/good users and questionnaire variables. p Values <.05 were considered statistically significant.
Number of users across minutes of CPAP use on all days. The distribution is positively skewed with 21% of CPAP users wearing CPAP ≤20 minutes. CPAP, continuous positive airway pressure.
RESULTS
The caregiver-reported SEMSA was completed by 161 consecutive caregivers and 70 patients who were old enough and developmentally able completed the self-report SEMSA. Adherence data were available for 141 patients at 1 month. Analyses were completed for 138 caregivers who completed the caregiver-reported SEMSA and 59 children who completed the self-report SEMSA and had 1-month adherence data available. Fisher’s exact tests revealed no statistical differences in race between caregivers and children included in analyses and those not included in analyses (p = .20 and p = .16, respectively).
Demographics are presented in Table 1. Children were aged 4 months to 18 years. SEMSA scores and adherence data are reported as median (IQR) as the distribution of these variables was not normal (Table 2). The distribution of the total CPAP usage on all days (in minutes) was skewed to the right (Figure 1). As expected, all adherence variables were highly correlated with each other, as were some health cognition variables (health cognition variables shown in Table 3). Average total minutes of use on all days correlated with percentage of days used (r = 0.82, p < .0001) and minutes of use on days that the device was used (r = 0.92, p < .0001). Percentage of days used and minutes of use on days used also correlated with each other (r = 0.70, p < .0001). Average total minutes of use on all days were used a priori as the primary outcome.
Demographic Characteristics of Patients (N = 141).
| Age (years), median (IQR) | 11.9 (7.9, 15.5) |
| *Infants and toddlers, N (%) | 4 (2.8) |
| *Preschool, n (%) | 6 (4.3) |
| *School aged, n (%) | 70 (49.6) |
| *Adolescents, n (%) | 61 (43.3) |
| Males, n (%) | 95 (67.4%) |
| Race, n (%) | |
| African-American | 70 (49.6%) |
| Caucasian | 53 (37.6%) |
| Other | 18 (12.8%) |
| Developmental disability, N (%) | 37 (26.2%) |
| AHI, median (IQR) | 13.8 (7.1, 29.7) |
| Diagnoses/characteristics associated with OSAS, N (%) | |
| Previous adenotonsillectomy | 102 (72.3%) |
| **Adenotonsillar hypertrophy | 16 (11.3) |
| Obesity | 77 (54.7%) |
| Genetic disorder | 10 (7.1%) |
| Craniofacial anomaly | 3 (2.1%) |
| CNS abnormality | 7 (5.0%) |
| Other comorbidity | 6 (4.3%) |
| Age (years), median (IQR) | 11.9 (7.9, 15.5) |
| *Infants and toddlers, N (%) | 4 (2.8) |
| *Preschool, n (%) | 6 (4.3) |
| *School aged, n (%) | 70 (49.6) |
| *Adolescents, n (%) | 61 (43.3) |
| Males, n (%) | 95 (67.4%) |
| Race, n (%) | |
| African-American | 70 (49.6%) |
| Caucasian | 53 (37.6%) |
| Other | 18 (12.8%) |
| Developmental disability, N (%) | 37 (26.2%) |
| AHI, median (IQR) | 13.8 (7.1, 29.7) |
| Diagnoses/characteristics associated with OSAS, N (%) | |
| Previous adenotonsillectomy | 102 (72.3%) |
| **Adenotonsillar hypertrophy | 16 (11.3) |
| Obesity | 77 (54.7%) |
| Genetic disorder | 10 (7.1%) |
| Craniofacial anomaly | 3 (2.1%) |
| CNS abnormality | 7 (5.0%) |
| Other comorbidity | 6 (4.3%) |
AHI, apnea–hypopnea index; IQR, interquartile range; OSAS, obstructive sleep apnea syndrome.
Some patients were classified in more than one category of diagnoses/characteristics.
*Children aged <3 years were categorized as infants and toddlers; preschoolers were considered aged 3–5.9 years, school aged were considered 6–12.9 years, and adolescents were considered 13–18 years.
**Parents declined surgery or patient had surgical contraindications.
Demographic Characteristics of Patients (N = 141).
| Age (years), median (IQR) | 11.9 (7.9, 15.5) |
| *Infants and toddlers, N (%) | 4 (2.8) |
| *Preschool, n (%) | 6 (4.3) |
| *School aged, n (%) | 70 (49.6) |
| *Adolescents, n (%) | 61 (43.3) |
| Males, n (%) | 95 (67.4%) |
| Race, n (%) | |
| African-American | 70 (49.6%) |
| Caucasian | 53 (37.6%) |
| Other | 18 (12.8%) |
| Developmental disability, N (%) | 37 (26.2%) |
| AHI, median (IQR) | 13.8 (7.1, 29.7) |
| Diagnoses/characteristics associated with OSAS, N (%) | |
| Previous adenotonsillectomy | 102 (72.3%) |
| **Adenotonsillar hypertrophy | 16 (11.3) |
| Obesity | 77 (54.7%) |
| Genetic disorder | 10 (7.1%) |
| Craniofacial anomaly | 3 (2.1%) |
| CNS abnormality | 7 (5.0%) |
| Other comorbidity | 6 (4.3%) |
| Age (years), median (IQR) | 11.9 (7.9, 15.5) |
| *Infants and toddlers, N (%) | 4 (2.8) |
| *Preschool, n (%) | 6 (4.3) |
| *School aged, n (%) | 70 (49.6) |
| *Adolescents, n (%) | 61 (43.3) |
| Males, n (%) | 95 (67.4%) |
| Race, n (%) | |
| African-American | 70 (49.6%) |
| Caucasian | 53 (37.6%) |
| Other | 18 (12.8%) |
| Developmental disability, N (%) | 37 (26.2%) |
| AHI, median (IQR) | 13.8 (7.1, 29.7) |
| Diagnoses/characteristics associated with OSAS, N (%) | |
| Previous adenotonsillectomy | 102 (72.3%) |
| **Adenotonsillar hypertrophy | 16 (11.3) |
| Obesity | 77 (54.7%) |
| Genetic disorder | 10 (7.1%) |
| Craniofacial anomaly | 3 (2.1%) |
| CNS abnormality | 7 (5.0%) |
| Other comorbidity | 6 (4.3%) |
AHI, apnea–hypopnea index; IQR, interquartile range; OSAS, obstructive sleep apnea syndrome.
Some patients were classified in more than one category of diagnoses/characteristics.
*Children aged <3 years were categorized as infants and toddlers; preschoolers were considered aged 3–5.9 years, school aged were considered 6–12.9 years, and adolescents were considered 13–18 years.
**Parents declined surgery or patient had surgical contraindications.
SEMSA and Adherence Descriptives.
| SEMSA | Median (IQR) |
|---|---|
| Caregiver-reported risk perception (N = 137) | 2.5 (2.0, 3.3) |
| Caregiver-reported outcome expectation (N = 137) | 3.3 (2.9, 3.9) |
| Caregiver-reported treatment self-efficacy (N = 138) | 2.9 (2.5, 3.7) |
| Patient-reported risk perception (N = 59) | 2.3 (1.5, 3.0) |
| Patient-reported outcome expectation (N = 59) | 2.9 (2.3, 3.4) |
| Patient-reported treatment self-efficacy (N = 59) | 2.7 (2.0, 3.6) |
| Adherence | |
| Mean total minutes of usage across all days (N = 141) | 176.4 (37.3, 350.2) |
| Percentage of days with usage at 1 month (N = 141) | 74.2 (35.4, 93.5) |
| Mean total minutes of usage on days used (N = 141) | 269.7 (95.9, 405.2) |
| SEMSA | Median (IQR) |
|---|---|
| Caregiver-reported risk perception (N = 137) | 2.5 (2.0, 3.3) |
| Caregiver-reported outcome expectation (N = 137) | 3.3 (2.9, 3.9) |
| Caregiver-reported treatment self-efficacy (N = 138) | 2.9 (2.5, 3.7) |
| Patient-reported risk perception (N = 59) | 2.3 (1.5, 3.0) |
| Patient-reported outcome expectation (N = 59) | 2.9 (2.3, 3.4) |
| Patient-reported treatment self-efficacy (N = 59) | 2.7 (2.0, 3.6) |
| Adherence | |
| Mean total minutes of usage across all days (N = 141) | 176.4 (37.3, 350.2) |
| Percentage of days with usage at 1 month (N = 141) | 74.2 (35.4, 93.5) |
| Mean total minutes of usage on days used (N = 141) | 269.7 (95.9, 405.2) |
IRQ, interquartile range; SEMSA, Self-Efficacy Measure for Sleep Apnea.
SEMSA and Adherence Descriptives.
| SEMSA | Median (IQR) |
|---|---|
| Caregiver-reported risk perception (N = 137) | 2.5 (2.0, 3.3) |
| Caregiver-reported outcome expectation (N = 137) | 3.3 (2.9, 3.9) |
| Caregiver-reported treatment self-efficacy (N = 138) | 2.9 (2.5, 3.7) |
| Patient-reported risk perception (N = 59) | 2.3 (1.5, 3.0) |
| Patient-reported outcome expectation (N = 59) | 2.9 (2.3, 3.4) |
| Patient-reported treatment self-efficacy (N = 59) | 2.7 (2.0, 3.6) |
| Adherence | |
| Mean total minutes of usage across all days (N = 141) | 176.4 (37.3, 350.2) |
| Percentage of days with usage at 1 month (N = 141) | 74.2 (35.4, 93.5) |
| Mean total minutes of usage on days used (N = 141) | 269.7 (95.9, 405.2) |
| SEMSA | Median (IQR) |
|---|---|
| Caregiver-reported risk perception (N = 137) | 2.5 (2.0, 3.3) |
| Caregiver-reported outcome expectation (N = 137) | 3.3 (2.9, 3.9) |
| Caregiver-reported treatment self-efficacy (N = 138) | 2.9 (2.5, 3.7) |
| Patient-reported risk perception (N = 59) | 2.3 (1.5, 3.0) |
| Patient-reported outcome expectation (N = 59) | 2.9 (2.3, 3.4) |
| Patient-reported treatment self-efficacy (N = 59) | 2.7 (2.0, 3.6) |
| Adherence | |
| Mean total minutes of usage across all days (N = 141) | 176.4 (37.3, 350.2) |
| Percentage of days with usage at 1 month (N = 141) | 74.2 (35.4, 93.5) |
| Mean total minutes of usage on days used (N = 141) | 269.7 (95.9, 405.2) |
IRQ, interquartile range; SEMSA, Self-Efficacy Measure for Sleep Apnea.
SEMSA Health Cognition Correlation Matrix.
| SEMSA Variable | Caregiver outcome expectations | Caregiver self-efficacy | Patient risk perception | Patient outcome expectations | Patient self-efficacy |
|---|---|---|---|---|---|
| Caregiver risk perception | 0.32** | 0.13 | 0.49** | 0.17 | 0.07 |
| Caregiver outcome expectations | 0.34** | 0.28* | 0.28* | 0.11 | |
| Caregiver self-efficacy | 0.16 | 0.27* | 0.27* | ||
| Patient risk perception | 0.35** | 0.11 | |||
| Patient outcome expectations | 0.51** |
| SEMSA Variable | Caregiver outcome expectations | Caregiver self-efficacy | Patient risk perception | Patient outcome expectations | Patient self-efficacy |
|---|---|---|---|---|---|
| Caregiver risk perception | 0.32** | 0.13 | 0.49** | 0.17 | 0.07 |
| Caregiver outcome expectations | 0.34** | 0.28* | 0.28* | 0.11 | |
| Caregiver self-efficacy | 0.16 | 0.27* | 0.27* | ||
| Patient risk perception | 0.35** | 0.11 | |||
| Patient outcome expectations | 0.51** |
*p < .05; **p < .01.
SEMSA Health Cognition Correlation Matrix.
| SEMSA Variable | Caregiver outcome expectations | Caregiver self-efficacy | Patient risk perception | Patient outcome expectations | Patient self-efficacy |
|---|---|---|---|---|---|
| Caregiver risk perception | 0.32** | 0.13 | 0.49** | 0.17 | 0.07 |
| Caregiver outcome expectations | 0.34** | 0.28* | 0.28* | 0.11 | |
| Caregiver self-efficacy | 0.16 | 0.27* | 0.27* | ||
| Patient risk perception | 0.35** | 0.11 | |||
| Patient outcome expectations | 0.51** |
| SEMSA Variable | Caregiver outcome expectations | Caregiver self-efficacy | Patient risk perception | Patient outcome expectations | Patient self-efficacy |
|---|---|---|---|---|---|
| Caregiver risk perception | 0.32** | 0.13 | 0.49** | 0.17 | 0.07 |
| Caregiver outcome expectations | 0.34** | 0.28* | 0.28* | 0.11 | |
| Caregiver self-efficacy | 0.16 | 0.27* | 0.27* | ||
| Patient risk perception | 0.35** | 0.11 | |||
| Patient outcome expectations | 0.51** |
*p < .05; **p < .01.
Primary Analyses
Unadjusted and adjusted regression results are presented in Table 4. According to the unadjusted OLS regression, the association between caregiver-reported self-efficacy and adherence was statistically significant (p < .01); mean total CPAP usage on all days is estimated to increase by 48.4 minutes when caregiver-reported self-efficacy increases by 1, with a 95% confidence interval (CI) of 13.4 minutes to 83.4 minutes. No other caregiver- or patient-reported health cognition variables were statistically significantly related to total minutes of CPAP use (Table 4). The adjusted linear regression showed a statistically significant interaction effect of sex on the association between caregiver-reported risk perception and adherence (p = .02). This indicated that when caregiver-reported risk perception increased by 1, the total usage on all days is estimated to increase by 31.7 minutes for male participants, whereas it was estimated to decrease by 36.2 minutes for female participants. No other interactions were statistically significant in this model.
Caregiver- and Patient-Reported Health Cognition Variables and Adherence.
| Regression for mean | Adherence (total minutes of use across all days) | ||||||
|---|---|---|---|---|---|---|---|
| Unadjusteda | Adjustedb | ||||||
| Estimates (95% CI) | p-value | p value for the interaction | |||||
| Age | Sex | Race | Obesity | Developmental disability | |||
| Caregiver-reported risk perception | 9.8 (−18.1, 37.7) | .49 | .30 | .02* | .84 | .63 | .66 |
| Caregiver-reported outcome expectations | 7.1 (−35.1, 49.3) | .74 | .12 | .82 | .13 | .12 | .80 |
| Caregiver-reported self-efficacy | 48.4 (13.4, 83.4) | .007* | .72 | .18 | .33 | .23 | .41 |
| Patient-reported risk perception | −21.2 (−63.2, 20.8) | .32 | .12 | .98 | .11 | .48 | N/A |
| Patient-reported outcome expectations | 30.3 (−28.0, 88.5) | .30 | .25 | .89 | .53 | .14 | N/A |
| Patient-reported self-efficacy | 38.3 (−5.2, 81.9) | .08 | .34 | .89 | .73 | .95 | N/A |
| Regression for mean | Adherence (total minutes of use across all days) | ||||||
|---|---|---|---|---|---|---|---|
| Unadjusteda | Adjustedb | ||||||
| Estimates (95% CI) | p-value | p value for the interaction | |||||
| Age | Sex | Race | Obesity | Developmental disability | |||
| Caregiver-reported risk perception | 9.8 (−18.1, 37.7) | .49 | .30 | .02* | .84 | .63 | .66 |
| Caregiver-reported outcome expectations | 7.1 (−35.1, 49.3) | .74 | .12 | .82 | .13 | .12 | .80 |
| Caregiver-reported self-efficacy | 48.4 (13.4, 83.4) | .007* | .72 | .18 | .33 | .23 | .41 |
| Patient-reported risk perception | −21.2 (−63.2, 20.8) | .32 | .12 | .98 | .11 | .48 | N/A |
| Patient-reported outcome expectations | 30.3 (−28.0, 88.5) | .30 | .25 | .89 | .53 | .14 | N/A |
| Patient-reported self-efficacy | 38.3 (−5.2, 81.9) | .08 | .34 | .89 | .73 | .95 | N/A |
*Statistically significant p < .05.
Adherence was fitted on each of the SEMSA variables both awithout adjustment and bwith adjustment for age, sex, race, obesity, and developmental disability (separately), including their interactions with SEMSA.
CI, confidence interval; SEMSA, Self-Efficacy Measure for Sleep Apnea.
Caregiver- and Patient-Reported Health Cognition Variables and Adherence.
| Regression for mean | Adherence (total minutes of use across all days) | ||||||
|---|---|---|---|---|---|---|---|
| Unadjusteda | Adjustedb | ||||||
| Estimates (95% CI) | p-value | p value for the interaction | |||||
| Age | Sex | Race | Obesity | Developmental disability | |||
| Caregiver-reported risk perception | 9.8 (−18.1, 37.7) | .49 | .30 | .02* | .84 | .63 | .66 |
| Caregiver-reported outcome expectations | 7.1 (−35.1, 49.3) | .74 | .12 | .82 | .13 | .12 | .80 |
| Caregiver-reported self-efficacy | 48.4 (13.4, 83.4) | .007* | .72 | .18 | .33 | .23 | .41 |
| Patient-reported risk perception | −21.2 (−63.2, 20.8) | .32 | .12 | .98 | .11 | .48 | N/A |
| Patient-reported outcome expectations | 30.3 (−28.0, 88.5) | .30 | .25 | .89 | .53 | .14 | N/A |
| Patient-reported self-efficacy | 38.3 (−5.2, 81.9) | .08 | .34 | .89 | .73 | .95 | N/A |
| Regression for mean | Adherence (total minutes of use across all days) | ||||||
|---|---|---|---|---|---|---|---|
| Unadjusteda | Adjustedb | ||||||
| Estimates (95% CI) | p-value | p value for the interaction | |||||
| Age | Sex | Race | Obesity | Developmental disability | |||
| Caregiver-reported risk perception | 9.8 (−18.1, 37.7) | .49 | .30 | .02* | .84 | .63 | .66 |
| Caregiver-reported outcome expectations | 7.1 (−35.1, 49.3) | .74 | .12 | .82 | .13 | .12 | .80 |
| Caregiver-reported self-efficacy | 48.4 (13.4, 83.4) | .007* | .72 | .18 | .33 | .23 | .41 |
| Patient-reported risk perception | −21.2 (−63.2, 20.8) | .32 | .12 | .98 | .11 | .48 | N/A |
| Patient-reported outcome expectations | 30.3 (−28.0, 88.5) | .30 | .25 | .89 | .53 | .14 | N/A |
| Patient-reported self-efficacy | 38.3 (−5.2, 81.9) | .08 | .34 | .89 | .73 | .95 | N/A |
*Statistically significant p < .05.
Adherence was fitted on each of the SEMSA variables both awithout adjustment and bwith adjustment for age, sex, race, obesity, and developmental disability (separately), including their interactions with SEMSA.
CI, confidence interval; SEMSA, Self-Efficacy Measure for Sleep Apnea.
Secondary Analyses
Quantile regression of the median was also performed given the skewed distribution; results were in the same direction as the OLS regression (Table 5). Visual review of data showed a clear decrease between the frequencies of children wearing CPAP < 20 minutes a night (21% of participants) compared to all other degrees of usage (Figure 1). Therefore, unadjusted logistic regression, dichotomized at 20 minutes, was performed to assess differences between minimal users compared to those with some CPAP use and revealed a trend for an association between wearing CPAP ≥ 20 minutes or < 20 minutes and SEMSA variables (p = .05). This trend was similarly observed when children with and without neurodevelopmental diagnoses were included in analyses (p = .05). Results for the caregiver-reported self-efficacy were consistent with the OLS and quantile regressions, suggesting a positive association with the CPAP adherence with an estimated odds ratio (OR) >1 (OR = 1.65, p = .06) and a 95% CI of 0.99 to 2.77.
Unadjusted Quantile Regression of Health Cognitions and Adherence.
| Quantile regression at the median | Adherence (total minutes of use across all days) | |
|---|---|---|
| Estimates (95% CI) | p-value | |
| Caregiver-reported risk perception | 25.3 (−24.8, 43.7) | .34 |
| Caregiver-reported outcome expectations | 15.2 (−70.0, 45.8) | .70 |
| Caregiver-reported self-efficacy | 49.8 (21.4, 99.3) | .06 |
| Patient-reported risk perception | −12.1 (−64.4, 5.2) | .75 |
| Patient-reported outcome expectations | 4.8 (−33.2, 120.3) | .93 |
| Patient-reported self-efficacy | 8.4 (5.0, 104.1) | .82 |
| Quantile regression at the median | Adherence (total minutes of use across all days) | |
|---|---|---|
| Estimates (95% CI) | p-value | |
| Caregiver-reported risk perception | 25.3 (−24.8, 43.7) | .34 |
| Caregiver-reported outcome expectations | 15.2 (−70.0, 45.8) | .70 |
| Caregiver-reported self-efficacy | 49.8 (21.4, 99.3) | .06 |
| Patient-reported risk perception | −12.1 (−64.4, 5.2) | .75 |
| Patient-reported outcome expectations | 4.8 (−33.2, 120.3) | .93 |
| Patient-reported self-efficacy | 8.4 (5.0, 104.1) | .82 |
CI, confidence interval.
Unadjusted Quantile Regression of Health Cognitions and Adherence.
| Quantile regression at the median | Adherence (total minutes of use across all days) | |
|---|---|---|
| Estimates (95% CI) | p-value | |
| Caregiver-reported risk perception | 25.3 (−24.8, 43.7) | .34 |
| Caregiver-reported outcome expectations | 15.2 (−70.0, 45.8) | .70 |
| Caregiver-reported self-efficacy | 49.8 (21.4, 99.3) | .06 |
| Patient-reported risk perception | −12.1 (−64.4, 5.2) | .75 |
| Patient-reported outcome expectations | 4.8 (−33.2, 120.3) | .93 |
| Patient-reported self-efficacy | 8.4 (5.0, 104.1) | .82 |
| Quantile regression at the median | Adherence (total minutes of use across all days) | |
|---|---|---|
| Estimates (95% CI) | p-value | |
| Caregiver-reported risk perception | 25.3 (−24.8, 43.7) | .34 |
| Caregiver-reported outcome expectations | 15.2 (−70.0, 45.8) | .70 |
| Caregiver-reported self-efficacy | 49.8 (21.4, 99.3) | .06 |
| Patient-reported risk perception | −12.1 (−64.4, 5.2) | .75 |
| Patient-reported outcome expectations | 4.8 (−33.2, 120.3) | .93 |
| Patient-reported self-efficacy | 8.4 (5.0, 104.1) | .82 |
CI, confidence interval.
DISCUSSION
This study shows that caregiver-reported self-efficacy is an important factor as it relates to the initiation of CPAP in youth. Although both caregiver- and patient-reported self-efficacy were examined as they relate to adherence, only caregiver self-efficacy emerged as a significant factor in short-term CPAP adherence. These results support the notion that caregivers and caregiver beliefs are an important consideration when prescribing CPAP in youth. While patient-reported perceptions were not found to be related to adherence in this study, it is important to study both patient and caregiver perceptions given the socioecological context in which CPAP is implemented, as well as the potential long-term need for CPAP. It may be that caregiver perception is most important at the initiation of CPAP, and over time, patient perception also becomes an important factor. Longitudinal studies of patient- and caregiver-reported health cognitions and CPAP use over time are needed.
These results also suggest that caregivers’ self-efficacy facilitates the initiation of health behaviors across childhood and adolescence, as age was not a significant covariate. While medical teams need to be sensitive to respecting the development of independence in youth, especially adolescents, these results suggest that strategies to improve self-efficacy in both caregivers and youth of all ages should be an important focus when implementing a complex medical intervention at home, such as CPAP.
Another novel aspect of this study is that the health belief cognitions were assessed following a semi-standardized educational and individualized behavioral plan regarding OSAS and CPAP treatment. The program was designed to improve caregiver and patient knowledge and self-efficacy. Self-efficacy has consistently been identified in adult populations as a predictor of CPAP adherence,7,13,14,16,17 but data are scarce in the pediatric population. For other chronic medical interventions implemented in pediatric patients, such as diabetes care, greater caregiver, and youth self-efficacy have been shown to be related to increased adherence.18–20 It is believed that promoting adaptive health cognitions may predict whether a person is likely to implement or engage in a particular health behavior, including implementation of CPAP.17,21 Future studies should examine caregiver- and patient-reported self-efficacy before and after the semi-standardized CPAP program, as well as after a trial of CPAP to better assess the benefits of the CPAP program, itself, and whether practice and implementation of CPAP in turn improves self-efficacy. If a family has difficulty implementing CPAP in the first month, self-efficacy may decrease, thereby making it more difficult to implement CPAP in the future. Longitudinal assessment of self-efficacy, cognitive behavioral intervention, and CPAP use is an important area of understudied research.
The current study also revealed that caregiver-reported risk perception was influenced by sex such that as caregiver risk perception increased, adherence increased for males but decreased for females. It has been reported that there are sex differences with regard to medical adherence in youth, with girls reporting less adherence than boys.22 It has also been shown that girls assume more responsibility for self-management of medications than boys.23 It is possible that while caregivers report similar levels of risk perception in girls and boys, girls may be more responsible for their own self-management and adherence suffers, whereas caregivers of boys continue to actively assist their child with the medical intervention providing the necessary support to adhere to medical recommendations. Further studies are needed to examine sex differences in adherence to medical interventions and caregiver perception of risk.
In contrast to a study in adults,21 this study did not demonstrate an association between caregiver- or patient-reported perceived risk and outcome expectations and adherence to CPAP at 1 month. Olsen et al. found that in adults, the greatest proportion of CPAP adherence was explained by higher outcome expectancies, greater functional limitations, and lower risk perception.21 One significant difference between the studies is that even though patients in both studies were CPAP naïve, participants in the Olsen study were provided with an explanation of their OSAS diagnosis and were presented with the recommended treatment of CPAP only. In the current study, caregivers and patients were provided a very detailed semi-standardized educational program, a physical demonstration on use of the equipment with the child, and an individualized behavioral plan. The current study employed a model more similar to Sawyer et al. where the SEMSA was administered following the detailed structured clinical visit and prior to CPAP use.24 The finding that self-efficacy is an important determinant of CPAP adherence at 1 month in the current study is similar to the findings of Sawyer et al., who showed that self-efficacy was associated with CPAP adherence in adults but in the context of knowledge of CPAP and treatment use.24 These results also support the value of a multidisciplinary approach that includes a behavioral health professional for the treatment of OSAS in youth with CPAP, in order to improve the belief and confidence in the ability of the patient and members of the family in implementing CPAP.
When examining interrelations among caregiver- and patient-reported health cognitions, we found that caregiver- and patient-reported risk perception and outcome expectation were moderately associated with one another, whereas the correlation for self-efficacy between caregiver and child was low to moderate. This finding may be the result of differences in these constructs, such that risk perception and outcome expectations are of an educational nature, whereas self-efficacy is an internal personal belief. Taken with the primary study results that caregiver self-efficacy is important in initiating CPAP, assessing and working to improve patients’ and caregivers’ level of self-efficacy may be more valuable than focusing on risk perceptions and outcome expectations of CPAP. However, additional studies are needed to further examine how these health-related cognitions are interrelated among caregivers and patients and how these may vary in their association with ongoing CPAP adherence over time.
Limitations
This study was part of a quality improvement initiative to improve the care of children and adolescents prescribed CPAP for the treatment of OSAS. Everyone prescribed CPAP received similar education. However, the behavioral plan was individually tailored to the patient and his or her family given the wide age range and large number of varying comorbid conditions of the patients, such as craniofacial anomalies, developmental disabilities, and genetic conditions. Therefore, there was variation in the process of the clinical visit, as well as a great deal of heterogeneity of patients starting CPAP, which could have impacted the health cognitions measured. However, each member of the clinical team (ie, physician, nurse, RT, and psychologist) did meet with all families and provided a “real world,” feasible CPAP training program, so the results found are likely generalizable. Further, this complex population is typical of children requiring CPAP.25,26
Maternal education level has been found to be related to adherence to CPAP4; however, this information was not collected in the current study. Future studies should examine this and other sociodemographic characteristics of caregivers in relation to health-related cognitions and adherence, as these efforts could help to inform interventions directed at increasing adherence on the basis of health-related beliefs.
In addition, most patients were still awaiting a titration polysomnogram at 1 month. It is our Center’s practice to initiate patients on lower pressures (4–6 cm H2O) for an adjustment period while awaiting their titration polysomnogram so they are able to successfully undergo titration. While the pressures that most children were on likely did not eliminate all events, the focus of this study is examining the initiation of CPAP. Longer term studies are needed to evaluate health cognitions on the maintenance of CPAP use and adherence to CPAP over time, after receiving therapeutic CPAP pressures that eliminate most/all events.
This study was not designed as an intervention study so it is unclear if ratings of self-efficacy, risk perception, and outcome expectations were improved following the CPAP training. Future research should explore health cognitions before and after CPAP training and whether changes in health belief cognitions improves initiation of CPAP in the short term. Finally, the constructs of self-efficacy, risk perception, and outcome expectations have multicollinearity, as demonstrated in this study and in another in adults24 and may contribute to differences in findings between cohorts and studies. As such, future replication studies examining these health belief cognitions are needed in larger samples. This study evaluated adherence at 1 month. Future studies are needed examining whether health cognitions, particularly caregiver self-efficacy, predict long-term CPAP use.
CONCLUSION
Adherence to CPAP has become an increasingly critical clinical concern given the growing number of youth prescribed CPAP. The results of this study indicate that employing strategies to improve self-efficacy when initiating CPAP with children, adolescents, and their families may promote CPAP adherence. These results also suggest that caregiver self-efficacy is a worthwhile measure to utilize in order to screen for the potential of nonadherence to the initiation of CPAP. Early identification of low caregiver-reported self-efficacy may make implementation of CPAP more effective because, once identified, providers can implement more adherence-promoting strategies early in the process. Further, these results provide a rationale to develop studies evaluating adherence-promoting interventions for the implementation of CPAP in youth that include strategies to improve caregiver-reported and patient-reported self-efficacy.
FUNDING
Study data were collected and managed using Research Electronic Data Capture (REDCap).
DISCLOSURE STATEMENT
TEW reports grants from TEVA, grants and other from Jazz Pharmaceutical, other from Philips Respironics, other from ResMed, and other from Nyxoah, outside the submitted work. CLM reports grants from NIH, outside the submitted work. No other authors have any other disclosures to report.
ACKNOWLEDGMENTS
We would like to thank the patients and caregivers for being active members of our quality improvement efforts. We would also like to thank Kristyna Greer, RN and Julianne Smith, MEd, RRT for assistance in downloading CPAP smartcard data. To obtain a copy of the Self-Efficacy Measure for Sleep Apnea, contact TEW at [email protected].
REFERENCES
Author notes
Address correspondence to: Melissa S. Xanthopoulos, PhD, Sleep Center, 9 NW 50, Children’s Hospital of Philadelphia, 34th and Civic Center Boulevard, Philadelphia, PA 19104, USA. Telephone: 215-590-4042; Fax: 267-426-9234; Email: [email protected]
INSTITUTION IN WHICH WORK WAS PERFORMED
Children’s Hospital of Philadelphia.
Comments