Correlation between preoperative quality-of-life (QOL) impact scores on the 18-item pediatric obstructive sleep apnea scale (OSA-18) and preoperative Child Behavior Checklist total problem T score classifications. Correlation is significant; log-likelihood ratio, P<.001.
Goldstein NA, Fatima M, Campbell TF, Rosenfeld RM. Child Behavior and Quality of Life Before and After Tonsillectomy and Adenoidectomy. Arch Otolaryngol Head Neck Surg. 2002;128(7):770-775. doi:10.1001/archotol.128.7.770
Copyright 2002 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2002
To determine the relationship between child behavior and quality of life before and after tonsillectomy and adenoidectomy by means of a standardized assessment of child behavior, the Child Behavior Checklist (CBCL), and a validated quality-of-life survey of pediatric obstructive sleep apnea, the OSA-18.
Hospital-based pediatric otolaryngology practice in a metropolitan area.
Sixty-four children (mean [SD] age, 5.8 [3.1] years; 36 boys, 28 girls) who underwent tonsillectomy and adenoidectomy for treatment of sleep-disordered breathing or recurrent tonsillitis.
Parents or caretakers completed the OSA-18 and the CBCL for ages 2 to 3 years or 4 to 18 years before surgery and 3 months postoperatively.
Main Outcome Measures
The OSA-18 mean survey scores and change scores, and the CBCL total problem T scores and change in total problem T scores.
The mean (SD) preoperative OSA-18 score was 3.9 (1.5) and change score was 2.3 (95% confidence interval, 1.9-2.7). The mean total problem score was 7.3 points lower after surgery (95% confidence interval, 4.9-9.7), indicating a significant decrease (P<.001, matched t test). The preoperative CBCL total problem score was consistent with abnormal behavior for 16 children (25%), but only 5 children (8%) scored in the abnormal range postoperatively (P = .03, log-likelihood ratio test). The OSA-18 preoperative mean survey score had fair to good correlation with the preoperative CBCL total problem T score (r = 0.50, P<.001, Pearson correlation), and the OSA-18 change score had fair to good correlation with the change in CBCL total problem T score (r = 0.54, P<.001, Pearson correlation).
Behavioral and emotional difficulties are found in children with sleep-disordered breathing before treatment and improve after intervention. Scores on a standardized measure of assessment of behavior demonstrate significant correlation with scores on a validated quality-of-life instrument.
SLEEP-DISORDERED BREATHING in children is viewed as a continuum of severity from partial obstruction of the upper airway producing snoring to increased upper airway resistance to continuous episodes of complete upper airway obstruction or obstructive apnea. The prevalence of pediatric obstructive sleep apnea syndrome (OSAS) has been estimated to be between 1% and 3% in preschool and school-aged children.1 Children with OSAS are usually brought to medical attention because of nighttime breathing difficulties. While behavioral, emotional, and neurocognitive difficulties have long been anecdotally associated with pediatric OSAS, recent reports using standardized behavioral assessments are supporting these observations.1- 7
The Child Behavior Checklist (CBCL) is a standardized measure of children's behavior.8,9 In a pilot study by our group of 36 children undergoing tonsillectomy and adenoidectomy (T & A) for chronic upper airway obstruction,3 10 (28%) scored in the abnormal range on the preoperative CBCL. The CBCL scores significantly improved after T & A (n = 15), although changes in classification (normal, borderline, or abnormal) were not statistically significant. Limitations of that study included a small sample size and lack of a control group.
The OSA-18 is an 18-item health-related quality-of-life (QOL) survey of pediatric OSAS that has been correlated with the respiratory disturbance index on nap polysomnography.10 Because of its ease of administration, reliability, validity, and responsiveness, it is a practical method of determination of the QOL impact from OSAS.10,11
We evaluated the impact of T & A on children's behavioral and emotional problems by means of the CBCL and the OSA-18. Surgery was performed for treatment of OSAS–upper airway obstruction, recurrent tonsillitis, or both. Correlation was sought between CBCL and OSA-18 scores. Changes in postoperative scores were assessed to evaluate longitudinal change in both behavior and QOL.
Parents of children aged 2 through 18 years who were scheduled for T & A or tonsillectomy were recruited from the private practice office of 3 pediatric otolaryngologists (including N.A.G. and R.M.R.) and the pediatric otolaryngology clinic of the Long Island College Hospital in Brooklyn, NY, from July 1, 1999, through December 31, 2000. Surgery was performed for treatment of OSAS–upper airway obstruction and/or recurrent tonsillitis. Children were diagnosed as having OSAS by a clinical history of nighttime snoring, apneic pauses, and struggling to breath, and by a physical examination documenting hypertrophic tonsils and adenoids. Polysomnography was not routinely performed. Children with craniofacial syndromes, neuromuscular diseases, behavioral disorders, learning disabilities, and psychiatric disease, and children whose parents did not read or understand English, were excluded. The protocol was approved by the institutional review board of the Long Island College Hospital, informed consent was obtained from the parents, and a convenience sample was recruited.
Parents or caretakers completed the OSA-18 and the CBCL for ages 2 through 3 years (CBCL/2-3) or ages 4 through 18 years (CBCL/4-18) before surgery and 3 months postoperatively. At entry, parents also completed a standard questionnaire that recorded demographic information, family socioeconomic status, and parental education level.
The OSA-18 consists of 18 survey items divided into 5 domains (sleep disturbance, physical symptoms, emotional symptoms, daytime functioning, and caregiver concerns). The 18 items are scored with a 7-point ordinal scale assessing frequency of specific symptoms scored as follows: none of the time, 1; hardly any of the time, 2; a little of the time, 3; some of the time, 4; a good bit of the time, 5; most of the time, 6; and all of the time, 7. The scores on each of the 18 items are summed to produce a total survey score that can range from 18 to 126. The OSA-18 total survey scores less than 60 suggest a small impact on health-related QOL, scores between 60 and 80 suggest a moderate impact, and scores greater than 80 suggest a large impact.10 A mean survey score and individual domain mean scores are also calculated. The OSA-18 change scores are calculated by subtracting the postoperative mean survey score and individual domain mean scores from the preoperative mean survey and individual domain mean scores. Change scores range from −7.0 to 7.0, with negative numbers indicating deterioration and positive numbers indicating improvement in QOL. Change scores of less than 0.5 indicate trivial change, 0.5 to 0.9 indicate small change, 1.0 to 1.4 indicate moderate change, and 1.5 or greater indicate large change.11
The CBCL/2-3 is a 100-item survey and the CBCL/4-18 is a 113-item survey of specific childhood behaviors. Each item is scored as follows: not true, 0; somewhat or sometimes true, 1; and very true or often true, 2. The CBCL/4-18 also has 35 questions regarding children's competencies in school, activities, and social contexts, which are scored on the basis of the amount and quality of participation for children aged 6 and older (total competence). Raw scores are converted to normalized T scores, which are compared with the scores of children in normative samples. Scores are then divided into normal (<95th percentile), borderline (≥95th percentile but <98th percentile), or abnormal (≥98th percentile) ranges.
The CBCL/2-3 is scored to obtain a total problem score, which provides a global index of the child's behavioral and emotional function. The responses can be further divided into scores for internalizing (anxious or depressed and withdrawn) and externalizing (aggressive behavior and destructive behavior) groupings, and scores for the individual syndrome scales (anxious or depressed, withdrawn, sleep problems, somatic problems, aggressive behavior, and destructive behavior). The CBCL/4-18 is scored to obtain a total problem score, scores for internalizing (withdrawn, somatic complaints, anxious or depressed) and externalizing (delinquent behavior and aggressive behavior) groupings, and scores for the individual syndrome scales (withdrawn, somatic complaints, anxious or depressed, social problems, thought problems, attention problems, delinquent behavior, and aggressive behavior). The CBCL/4-18 total competence score is reported separately and not included in the total problem score calculation.
To determine the relationship between behavior and QOL, we included children undergoing surgery for treatment of recurrent infection without sleep-disordered breathing in all our analyses. This ensured a distribution of children with varying degrees of sleep-disordered breathing in the study sample.
Comparison of preoperative and postoperative CBCL T scores was performed with a matched, 2-tailed t test, and preoperative and postoperative T score classifications (normal, borderline, and abnormal) were compared by the log-likelihood ratio test. Change in impact of health-related QOL based on OSA-18 scores was compared by the χ2 test. The preoperative total problem T score categories were compared with the preoperative OSA-18 QOL impact scores by means of the log-likelihood ratio test. Pearson correlation was used to compare the mean preoperative OSA-18 survey scores and change scores with the CBCL preoperative total problem, internalizing, and externalizing T scores, and changes in T scores postoperatively. Analysis of variance was used to compare the CBCL T scores and OSA-18 mean survey and change scores with the parent's education level, family income, and patient sex. A linear regression model was used to compare the preoperative OSA-18 mean survey score and parent education level as predictors of the CBCL total problem T score. P<.05 was considered statistically significant.12
Parents of 133 children were offered participation in the protocol, and 71 agreed to participate. Complete preoperative and postoperative surveys were available for 64 children who composed the study population. Patient demographics and indication for surgery are presented in Table 1. The male-female ratio was 1.3:1, and the mean patient age was 5.8 years (range, 2.0-18.1 years). Fifty parents (83%) who completed the surveys had graduated from high school and 12 (20%) had graduated from college. Most families were middle or upper middle class (annual family income between $21 000 and $100 000).13 Surgery was performed for the treatment of OSAS–upper airway obstruction in 39 children (61%), recurrent tonsillitis in 10 (16%), and both obstruction and tonsillitis in 15 (23%). The majority of children with upper airway obstruction were diagnosed by clinical assessment, as only 5 (8%) had preoperative polysomnograms (mean respiratory disturbance index, 25.2; range, 3.9-74.5).
The OSA-18 preoperative and change scores are presented in Table 2. Sleep disturbance, physical symptoms, and caregiver concerns were the highest-rated domains, followed by emotional symptoms and daytime function. Change scores (preoperative mean scores minus postoperative mean scores) for sleep disturbance, caregiver concerns, and physical symptoms indicate a large change in QOL (change score of ≥1.5), while change scores for emotional symptoms and daytime function indicate a moderate change in QOL (change score of 1.0-1.4). Preoperatively, the health-related QOL impact of the children's sleep-disordered breathing was small for 22 children (34%), moderate for 16 children (25%), and large for 26 children (41%) (Table 3). A highly significant change (P<.001) was found in QOL postoperatively, where a small impact was found in 63 children (98%) and a moderate impact was found in 1 child (2%).
Mean preoperative CBCL T scores and change in T scores for the individual syndrome scales, internalizing and externalizing groupings, and the total problem score are presented in Table 4. Sample size varied because some of the syndrome scales are applicable to only the CBCL/2-3 or the CBCL/4-18, the total competence items are scored only for children aged 6 years and older, and 3 CBCLs had incomplete responses for the total competence items and were excluded from the analysis. The mean total problem score was 7.3 points lower after surgery, resulting in a statistically significant decrease (P<.001). Scores were significantly lower after surgery for the internalizing and externalizing subgroups and all but 2 of the individual syndrome scales (total competence and social problems). Preoperative and postoperative T score classifications are presented in Table 5. Although there were trends toward improvement, changes in classification were only significant for the total problem score and the individual problem scales of anxious or depressed, thought problems, and sleep problems.
Since a number of socioeconomic variables have been shown to have a significant influence on children's development,14 we used analysis of variance to compare parent's education level, family income, and patient sex with CBCL and OSA-18 results. There was no significant relationship between the 3 sociodemographic variables and the preoperative OSA-18 mean survey scores and change scores. There was also no significant relationship between family income and sex and the preoperative CBCL total problem, internalizing, and externalizing T scores, although the relationship between parent education level and the total problem and externalizing T scores approached significance (Table 6).
The correlation between the preoperative OSA-18 health-related QOL impact scores and the preoperative CBCL total problem score classifications is presented in Figure 1. All 22 children with an OSA-18 survey score indicating a small impact on QOL also had a CBCL total problem score in the normal range. Of the 16 children with a survey score indicating a moderate impact on QOL, 4 (25%) had a total problem score in the abnormal range. Of the 26 children with a survey score indicating a large impact on QOL, 12 (46%) had a total problem score in the abnormal range. These results are highly significant (P<.001).
The OSA-18 preoperative mean survey scores had fair to good correlation with the preoperative CBCL total problem scores (r = 0.50, P<.001) and the internalizing (r = 0.43, P<.001) and externalizing (r = 0.43, P<.001) subgroups. The OSA-18 change scores had fair to good correlation with the change in CBCL total problem scores (r = 0.54, P<.001) and the internalizing (r = 0.50, P<.001) and externalizing (r = 0.49, P<.001) subgroups. About 25% of the variations in one score could be explained by variations in the other.
Because the influence of parent education level on CBCL scores approached significance by univariate analysis, a linear regression model was constructed to compare the preoperative OSA-18 mean survey score and parent's education level as predictors of the total problem T score (Table 7). Although parents who were both high school and college graduates reported significantly lower total problem CBCL T scores than parents who did not finish high school, the mean OSA-18 score was a more significant predictor of CBCL results (P = .01 vs P<.001).
Our study population of 64 children undergoing T & A predominantly for the treatment of sleep disordered-breathing (84%) demonstrated a high prevalence of behavioral and emotional (25%) problems. In addition, CBCL T scores for the total problem score and the majority of the individual domains significantly improved after T & A. Changes in classification (normal, borderline, or abnormal) were also statistically significant for the total problem score and some of the individual domains. Only 8% of the children scored in the abnormal range for the total problem score postoperatively. The health-related QOL impact of sleep-disordered breathing was large (41%) or moderate (25%) for our study population, and significantly improved postoperatively (66% vs 2%). The relationship between behavior and QOL was both static and dynamic, since both the baseline and the change scores demonstrated significant correlation.
The CBCL results are similar to the results obtained in the pilot study by our group of 36 children undergoing T & A for treatment of chronic upper airway obstruction.3 Preoperatively, the total problem score was abnormal for 28% of the patients as compared with 25% in the present study. Individual syndrome scales most affected in both studies were total competence, sleep problems, and somatic complaints. The mean total problem score was 7.5 points lower 3 months postoperatively in the previous study (n = 15) as compared with 7.3 points in the present study. Although changes in classification were not significant in the previous study after surgery, we were able to demonstrate significant postoperative changes in the total problem score, anxious or depressed, thought problems, and sleep problems in the present study, which may be attributed to our larger sample size (N = 64).
Sociodemographic variables including parental education, family income, and race or ethnicity have been shown to influence children's developmental performance.14 Boys also tend to score higher on the CBCL than girls.8,9 Only parental education level influenced the CBCL results, as parents who had not graduated from high school reported higher CBCL scores for their children than did parents who had graduated from high school or from college. The relationship between OSA-18 QOL scores and CBCL scores, however, remained highly significant when adjusted for the confounding effect of parent education (Table 7). Together, the OSA-18 scores and parent education predicted 34% of the variance in CBCL scores.
Although widely cited as a common complication of pediatric OSAS, behavioral disturbances have mostly been inferred from parental reports of children's behavior before and after treatment. A few recent studies using standardized assessments have identified the presence of inattention, hyperactivity, memory problems, learning difficulties, and poor school performance in children with sleep-disordered breathing.1- 4,7 Our study also identifies thought problems and attention problems in children with sleep-disordered breathing, but hyperactivity is not a problem scale on the CBCL. Although the physiological mechanism by which OSAS produces its neurocognitive and behavioral effects is unknown, disrupted sleep patterns and hypoxia are considered integral to their development.15,16
The OSA-18 is the only QOL survey that has been validated as a discriminative measure of severity of pediatric sleep-disordered breathing and an evaluative measure of longitudinal change. The preoperative mean survey score and change score for our 64 patients were even greater than the scores obtained in the study performed to validate the responsiveness of the survey.11 In this previous study of 69 children undergoing tonsillectomy or adenoidectomy for the treatment of sleep-disordered breathing (78%) or recurrent tonsillitis (22%), the mean survey score was 3.1. The mean change score for the children with clinical evaluations suggestive of sleep-disordered breathing was 1.14. Although not an objective measure of the severity of sleep-disordered breathing, such as polysomnography, the OSA-18, like other health-related QOL surveys, focuses on the physical problems, functional limitations, and emotional consequences of OSAS. The OSA-18 is more comprehensive in its scope than polysomnography.
No child with an OSA-18 survey score demonstrating a small impact on QOL scored in the abnormal range on the CBCL total problem score. In contrast, 25% of children with an OSA-18 survey score demonstrating a moderate impact on QOL and 46% of children with a score demonstrating a large impact scored in the abnormal range. Fair to good correlation was found between the OSA-18 QOL and CBCL preoperative and change scores. Four items (22%) on the OSA-18 that evaluate the frequency of mood swings, aggression or hyperactivity, discipline, and inattention overlap with domains of the CBCL. Despite this overlap, the 2 survey instruments evaluate separate but related consequences of sleep-disordered breathing.
This is, to our knowledge, the first prospective study of pediatric obstructive sleep apnea correlating children's behavioral disturbances with QOL and documenting improvements in both after treatment using validated assessments. The major weakness is that "effectiveness" or "efficacy" cannot be inferred, since the study did not have a control group. Although improvements in both behavior and QOL occurred, we cannot be certain that they were the result of T & A.
Our study provides further evidence that behavioral and emotional difficulties are present in children with sleep-disordered breathing and improve after treatment. Scores on our standardized assessment of behavior demonstrated significant correlation with scores on a validated QOL instrument. Additional work is needed to define the precise spectrum of behavioral abnormalities, elucidate their pathophysiologic mechanism, and provide diagnostic clues to allow their early recognition.
Accepted for publication December 17, 2001.
This study was supported by a 1999 Research Award from the American Society of Pediatric Otolaryngology (Dr Goldstein).
This study was presented at the American Society of Pediatric Otolaryngology 16th Annual Meeting, Scottsdale, Ariz, May 10, 2001.
We thank Kathy Rivera for assistance with data collection.
Corresponding author and reprints: Nira A. Goldstein, MD, Department of Otolaryngology, State University of New York Downstate Medical Center, 450 Clarkson Ave, Box 126, Brooklyn, NY 11203-2098 (e-mail: firstname.lastname@example.org).