Objective. To investigate health care utilization of children with obstructive sleep apnea syndrome (OSAS) 1 year after adenotonsillectomy (T&A).
Methods. A longitudinal, case-controlled, prospective study was conducted at Clalit Health Care Services (CHS), a health maintenance organization in the southern region of Israel. We defined 3 groups of children: 1) children who had OSAS and were treated with T&A (n = 130); 2) children who had OSAS and did not undergo surgery (n = 90); and 3) control subjects who were matched by age, sex, and area of residency (n = 520) and randomly selected from the CHS database. OSAS was verified with polysomnography studies in all patients. Indices of health care utilization were analyzed 1 year before and 1 year after T&A. Medical records in the emergency department and during hospitalization were reviewed for diagnosis before the polysomnography diagnosis.
Results. Mean age of all children with OSAS was 5.6 ± 3.6 years. Total annual health care costs were reduced by one third in children who underwent T&A, whereas there was no change in the control and untreated OSAS groups. T&A was associated with a 60% reduction in the number of new admissions, 39% reduction in emergency department visits, 47% reduction in the number of consultations, and 22% reduction in costs for prescribed drugs. In group 2, the total costs were similar in years 1 and 2.
Conclusions. T&A significantly reduces health care utilization in children with OSAS. Untreated children with moderate and severe OSAS will continue to consume high levels of health care resources. Increased morbidity among children with OSAS is mainly related to upper respiratory tract infections.
Obstructive sleep apnea syndrome (OSAS) in children may be associated with growth failure, cor pulmonale, secondary enuresis, behavioral and neurocognitive problems, and poor school achievement.1–8 The prevalence of OSAS in children is estimated as 1% to 3%.1,9 The most common cause of OSAS in children is adenotonsillar hypertrophy; thus, adenotonsillectomy (T&A) is the treatment of choice.10 After T&A, the majority of children with OSAS will have both symptomatic and polysomnographic (PSG) improvement2,3,11 and will also show improvement in right ventricular function,6 growth,3,12–14 quality of life, and behavior.15
Little is known about the effects of OSAS and its treatment on health care utilization in children. We recently reported a 2.3-fold increase in health care utilization among children with OSAS 1 year before PSG diagnosis.16 Adults with OSAS are heavy consumers of health care services17–22 years before their OSAS is diagnosed in a sleep laboratory. Treatment with continuous positive airway pressure reduces health care utilization in adults21,23,24 and improves quality-adjusted life years25 and quality of life.26
In the present study, we investigated the effect of T&A in children with OSAS and on health care consumption 1 year before and after PSG diagnosis. In addition, we explored the comorbidity that led to the high health care consumption in children with OSAS 1 year before PSG diagnosis. We hypothesized that treating OSAS in children with T&A would result in a reduction in health care utilization.
This was a longitudinal, case-control, prospective study.
All children are members of Clalit Health Care Services (CHS), the largest health maintenance organization (HMO) in Israel. All participating subjects are inhabitants of the southern region of Israel, which includes ∼450 000 CHS enrollees. All families of included subjects had been permanent residents of the region for at least 3 years before study initiation. We compared health care utilization between “typical” OSAS and “typical” control patients. We excluded, in all groups, children who were younger than 1 year and those with additional chronic morbidity and subjects who exhibited extreme consumption of health care services, ie, >10 times the mean values of their group.
The study groups included 258 children who were aged 1 to 18 years and were referred to the Sleep-Wake Disorder Unit in the Soroka University Medical Center in Beer-Sheva for the evaluation of OSAS. Children were recruited consecutively during the period of August 1998 through August 2000. Thirty-eight children were found to have normal PSG findings and did not require T&A.
We defined 2 groups of children. The records of each group were abstracted for a total of 2 years. Group 1 comprised 130 children who had OSAS and underwent T&A (Table 1). Data were extracted for 1 year before PSG diagnosis (year 1) and 1 year postdiagnosis (year 2). However, to avoid confounding medical utilization data with perisurgical morbidity, we allowed a gap of 3 months postsurgery before initiating the review for year 2. The median duration of time between PSG diagnosis and T&A surgery in our region is 4 months (range: 1–10 months). In 1 child, 18 months’ duration was noted. This finding had negligible effect on the results. Group 2 comprised children who had OSAS and did not undergo surgery (n = 90). On conclusion of the study, a telephone survey was conducted among the parents of group 2 to understand the lack of adherence with the recommended therapy. Thirty-four of this group had moderate to severe OSAS and did not undergo T&A. No other therapy was provided. Parents’ explanations were mainly related to the low level of awareness of the benefits of early T&A intervention, and 15 parents refused therapy. Data were abstracted for the year before (year 1) and after PSG (year 2) diagnosis of OSAS, respectively (Table 1).
Control Group (Group 3)
Data were extracted for 2 years consecutively (n = 520). The first year (year 1) was the control for the “prediagnosis” year, and the second year (year 2) was the control for the “postdiagnosis” year of the other groups. This group was matched to patients who had OSAS and underwent T&A by age, sex, and area of residency (to adjust for socioeconomic factors). Four control subjects with no evidence for concomitant chronic disease16,20 were randomly matched (case controlled) to each of the OSAS patients (using Microsoft Access) from the regional CHS database, which contains 126 357 enrollees. We confirmed that the control group represents the population of normal children by verifying that they did not receive any long-term medication. None of the control subjects underwent a PSG study; we did not exclude snorers from this group.
The Sleep-Wake Disorder Unit in the Soroka University Medical Center is the only such center for children in the Negev. All children underwent nocturnal PSG monitoring (Sensor Medics Inc, Yorba Linda, CA).27 Subjects reported to the sleep laboratory at 8:30 PM and were discharged at 7:30 the following morning. They were encouraged to maintain their usual daily routine and to take medications as usual. Overnight PSG monitoring was performed as follows: 2 silver-silverchloride electroencephalographic electrodes filled with electrolyte were applied to the C3 and C4 locations, and reference electrodes were attached behind the ears in the left (A1) and right (A2) mastoid areas. Two electromyographic electrodes were applied over the submental muscles. Two electro-oculographic electrodes were applied 1 cm above the outer canthus of 1 eye and 1 cm below the outer canthus of the other eye. The montage arrangement for PSG reading consisted of C3A2 and O2A1, 2 electro-oculographic, and electrocardiogram (modified V2 lead). Nasal airflow was monitored by a pressure transducer (RespSponse, Synectics Sleep Inc, Stockholm, Sweden); thoracic and abdominal movements were monitored by strain-gauge electrodes; hemoglobin oxygen saturation was monitored by pulse oximetry (Ohmeda 4700, Temecula, CA). Sleep/wake sleep stages,28 arousals, and awakenings were scored as recommended, with the appropriate modifications for children3,29,30 and obstructive apneic and hypopneic events.9 The Ethics Committee of Ben-Gurion University of the Negev approved this project.
Health Care Utilization
The CHS billing system records contain >98% of all medical expenditures of the HMO. All costs were collected for each child in the 12-month period before the PSG evaluation and 12 months after T&A. The costs of PSG were not included in the study. Cost indicators include number of new admissions and hospitalization days, number of pediatric “day hospital” visits (<24-hour admission), number of emergency department visits, number of (new and repeated) visits to specialists, and drugs prescribed (type and cost for each drug). All prescribed drugs are categorized according to World Health Organization recommendations.31 For the purpose of this study, we defined hospital visits as the sum of new admissions and visits in the pediatric emergency departments; the registry of these parameters does not overlap. We defined “total annual cost” as the sum of the costs for all indicators. Costs are expressed as the mean per patient per year in $US according to the price list published by the Israeli Ministry of Health in January 2002. The exchange rate was NIS 4.2 per $1 US. Values were adjusted for inflation.
Relationship Between Respiratory Disturbance Index and the Effect of Surgery on Health Care Utilization
To investigate the effect of apnea severity and treatment on total annual costs, we calculated the difference in total annual costs between year 1 and year 2. A positive value indicates savings in health care consumption, whereas a negative value indicates an increase in health care consumption in year 2. Differences in health care consumption, in $US, were evaluated according to OSAS severity. For the purpose of this study, we arbitrarily defined 3 levels of respiratory disturbance index (RDI) severity according to the percentile of the distribution of the RDI results: mild RDI, 1 to 3.6 events/hour (33.3%); moderate RDI, 3.6 to 7.9 events/hour (33.3%–66.6%); and severe RDI, > 8 events/hour (66.6%–100%).
RDI was calculated as the number of respiratory events (apnea/hypopnea) per hour of sleep. Obstructive apneic and hypopneic events were scored according to the recommended pediatric criteria of the American Thoracic Society.9 Obstructive apnea was defined as paradoxic breathing for at least 2 respiratory cycles with complete cessation of nasal airflow (airflow reduction of at least 80%). Obstructive hypopnea was scored when the paradoxic breathing was accompanied by a reduction of at least 50% in airflow, resulting in oxygen desaturation of at least 4%.
Because Soroka University Medical Center is the only referral hospital in the region, we found 99% of all medical records needed. These medical records reflect all information regarding the emergency department visits and hospitalization. The medical information is classified according to the International Classification of Diseases, Ninth Revision. Two pediatricians reviewed the records. For the purpose of this study, they classified them into 3 categories: 1) upper respiratory infections that could be related to large tonsils and adenoids (otitis media, tonsillitis, laryngitis, croup); 2) lower respiratory diseases such as pneumonia, bronchiolitis, and asthma; and 3) other.
Cost data were analyzed according to previous methods16,32 using SPSS software. Three groups of children were analyzed: the control group (control), the T&A group (T&A), and the untreated OSAS group (untreated). One-way analysis of variance was used to determine differences in age, RDI, percentage of sleep time with saturation <90%, total sleep time, and arousal index (number of arousal events per hour of sleep). Two-way analysis of variance was used to determine the significance of total annual costs between groups 1 and 2 and apnea severity. The χ2 test was used to confirm “population match” between the control and OSAS populations in relation to sex. We calculated average admission and emergency department visits for each of the study groups to estimate costs. The Mann-Whitney test was used to determine statistical significance of cost elements between groups. The effect of treatment or time on all cost parameters was analyzed using the Wilcoxon signed rank test, as the distribution was nonparametric. Data were presented as mean ± standard deviation for all sleep parameters (Table 1) and as mean ± standard error of the mean for costs. The null hypothesis was rejected at the 5% level.
A total of 220 children who had OSAS and a mean age of 5.6 ± 3.6 years were recommended for T&A, 130 of whom (mean age: 5.1 ± 3.2 years; male/female: 51/79) underwent the procedure. Ninety children (mean age: 6.5 ± 4.1 year; P < .01; male/female: 29/61) did not undergo T&A. The control group included 520 children who were matched (χ2 test) by age and sex to the T&A group (mean age: 5.1 ± 3.2 years; male/female: 204/316).
There were no socioeconomic and demographic differences between groups 1 and 2. Both groups had an average RDI of 9.7 ± 8.1 and 8.1 ± 10.8 events/hour, respectively. As a group, children with OSAS did not present evidence of significant nocturnal hypoxemia but demonstrated a significantly higher number of arousals compared with control subjects.27 The arousals were more severe in the T&A group.
Review of the Medical Records
The average visit rate per child was ∼1.4 visits per year and was similar between groups and years. Table 2 summarizes the medical diagnosis made during hospital visits in groups 1 and 2, between years 1 and 2. The most notable clinical effect of T&A was 50% reduction in upper respiratory infections between years 1 and 2 in group 1 compared with group 2. In the lower respiratory and other diseases, the clinical effect was <10%.
Health Care Utilization
Health care costs among children with OSAS in year 1 were 2.5-fold (P < .001) higher than in the control group. Approximately 75% of the total annual cost of the OSAS and control groups in year 1 was attributable to 3 elements: new admissions, visits to the emergency department, and consultations (Table 3). T&A reduced total annual costs by 32.5% (P < .0004) during year 2 after the intervention (Fig 1). During the same period, total annual costs did not change significantly in the control and untreated OSAS groups.
Relationship Between RDI and the Effect of Surgery on Health Care Utilization
In group 1, a reduction in total annual costs of $130 per child was noted in the moderate and severe subgroups (n = 107). In group 2, minimal change in total annual cost was noted in the mild and moderate subgroups (n = 68). However, in the severe subgroup (n = 22), a marked elevation of $287 per child in total annual costs was noted. Two-way analysis of variance revealed that the effect of groups had a trend toward reduction of total annual costs (P < .1). The interaction between RDI and groups 1 and 2 was statistically significant (P < .01; Fig 2).
Table 3 summarizes the health care utilization during year 1 compared with year 2. Children who did and did not undergo T&A had greater health care utilization in year 1, compared with the control group (P < .001). In the untreated OSAS group, the number of admissions increased 3.1 (P < .02) in year 2. Approximately 1 specialist visit per year was documented in the control group in year 1. However, 3.6 (P < .001) and 4.4 visits per child to specialists were documented in the T&A and untreated groups, respectively. The most common reason for consultations in both OSAS groups in year 1 included otolaryngology surgeons in 43% of the children, pediatric pulmonologists in 20%, and ophthalmologists in 11%. Other consultations included neurologists, cardiologists, dermatologists, and orthopedic surgeons with <10% for each specialty. In the control group, 18% of the children visited otolaryngology surgeons, 5% visited pediatric pulmonologists, and 18% visited ophthalmologists. All other specialists had 1% to 5% referrals per specialty. One year after diagnosis, the T&A group had a significantly lower rate of visits to otolaryngology surgeons and pediatric pulmonologists: 32% and 11%, respectively. Other types of consultations were not affected by the T&A. In both the control and untreated groups, there was no significant change in the number of any of the consultations 1 year after diagnosis.
Ten percent to 12% of all annual costs in all groups were related to prescribed drugs. Table 4 summarizes the pharmacologic groups that accounted for approximately two thirds of the cost for prescribed drugs. Both the T&A and untreated OSAS groups had >3-fold increases (P < .001) in the utilization of prescribed drugs in year 1. In year 2, there was a significant reduction of ∼25% in the cost for drugs only in the T&A group (P < .006); no significant change was found in the control and untreated OSAS groups.
We previously reported16 that children who have OSAS up to 8 years of age are heavy consumers of health care resources. The main elements that compose this high consumption of health services are hospitalizations, visits to the emergency department, referrals to specialists, and drugs issued. Increased morbidity among children with OSAS is related mainly to upper respiratory tract infections.
The pathophysiology of OSAS in children is related to a combination of anatomic narrowing and neuromuscular function.2 OSAS results from the relative size and structure of the upper airway components, rather than from the absolute size of the adenotonsillar tissue. The peak prevalence of childhood OSAS occurs at 2 to 8 years, which is the age when tonsils and adenoids are the largest in relation to the site of collapse.33 The most common treatment for children with OSAS is T&A, which is associated with symptom resolution.10 In the present study, removal of the adenotonsillar tissue resulted in a one third reduction of total annual costs 1 year after the T&A. In otherwise healthy children who have adenotonsillar hypertrophy and undergo T&A, PSG resolution occurs in 75% to 100% (for a review see 1,34). Perhaps our findings of partial reduction of costs after T&A are additional support for the notion that the surgery per se does not provide full recovery in the short term to all children with OSAS. These children may need additional therapy.1,34 Can the remaining relatively elevated medical expenses after T&A be blamed solely on persisting OSAS? In the current study, children from group 1 were not evaluated with PSG study after T&A. Because T&A is the therapy of choice, we assumed that most of the children in group 1, in contrast to group 2, were cured and had an improved quality of life.15 However, we recently found35 that RDI during rapid eye movement sleep did not return to normal values 4 to 9 months after surgery in 36 “typical” children who had OSAS and underwent T&A. Thus, residual OSAS after T&A, at least in rapid eye movement sleep, may provide some explanation as to the remaining high cost of health care utilization. After treatment, children are less sick with upper respiratory illness and have non–OSAS-dependent illnesses that persist postoperatively (Table 2). Perhaps the removal of chronically sick tonsillar tissue causing nighttime airway obstruction led to a reduced tendency for airway infections. The explanation of these results requires additional studies.
Previously, we found that health care utilization is affected by age. The consumption of health care services was greater in children ≤5 years and in direct correlation to the severity of OSAS. In patients from group 2, health care utilization was reduced by one third comparing years 1 and 2. This effect is 3 times more than the effect of age per se during the first 8 years of life.16 Thus, we can conclude that the reduction of costs is attributable to the therapy.
Both the T&A and untreated groups had similar health care utilization 1 year before PSG diagnosis. This finding is probably explained by the fact that both the T&A and untreated OSAS groups were similar with respect to OSAS severity (Table 1). The results of the present study support our previous results showing a correlation between the severity of OSAS and total annual health expenditures. This parameter was found to affect health care expenditures 1 year before PSG study. The T&A group had more arousals per hour of sleep compared with the untreated group (P < .01); however, this parameter of sleep disruption was not found to influence medical costs.16
We found that it is most beneficial, in terms of annual savings, to treat children who have a diagnosis of severe OSAS (Fig 2). In patients with mild and moderate OSAS, no significant savings were found. Therefore, when treating children with OSAS, the total resulting savings is attributable to the treatment of children with severe OSAS. More studies are needed to establish the cost and benefits in patients with mild and moderate OSAS.
To our knowledge, this is the first report to document significant reduction in health care utilization among children who underwent T&A. Changes in health care consumption reflect changes in health status. Therefore, the reduction in the number of hospital admissions, emergency department visits, consultations, and prescription drugs reflects the effect of OSAS treatment on health status. In the untreated group, no significant change in health status was found in all parameters except for significant elevation of new admissions. Previous reports on adults with OSAS have demonstrated reductions in health care utilization in patients who have OSAS and are treated with continuous positive airway pressure.22,23
On conclusion of the study, a telephone survey was conducted among parents of children in group 2 to understand the lack of adherence with the recommended therapy. We found that in 1 case, the duration between PSG diagnosis and T&A surgery was >1 year, regardless of OSAS severity. Parents’ explanations were mainly related to the low level of awareness among patients, physicians, and administrators of the benefits of early T&A intervention. In 15 cases, lack of adherence was attributable to refusal of T&A. The lack of adherence was not associated with approval for therapy from the HMO, because in our health care system, T&A is provided freely to all citizens as part of the Israeli National Health Care Law. Our results support the need for surgery as soon as possible when recommended, mainly in the moderate and severe cases. A higher level of awareness among physicians and parents can improve adherence with the recommendation for surgery.
Previous reports have described the CHS economic database for the southern part of Israel in detail.16,36,37 The information presented in this study reflects the “true” consumption of health care resources by children with OSAS. The relevant medical information regarding patients with OSAS is stored in the only Sleep-Wake Disorder Center in the study region; CHS uses a single billing system; physicians are paid a capitation fee once every 3 months per patient and therefore do not have any economic incentive to increase consumption of services. Thus, it is reasonable to conclude that the presented costs for health care services are the minimum annual expenditures for children with OSAS, because none of the children investigated had concomitant diseases,34 which may result in even higher health care expenditures.
Our findings show that T&A in children, mainly with severe OSAS, results in a significant reduction in health care utilization. Most of the health care resources consumed before PSG diagnosis are attributable to upper respiratory diseases, which probably contribute to the development of nighttime upper airway obstruction. Early intervention as a result of to greater awareness of all parties is of significant health benefit. Additional research is needed to determine the long-term effect of T&A on OSAS severity, health care consumption, and morbidity.
- Received December 2, 2002.
- Accepted June 3, 2003.
- Reprint requests to (A. Tarasiuk) Department of Physiology, Faculty of Health Sciences, Ben-Gurion University of the Negev, PO Box 105, Beer-Sheva 84105, Israel. E-mail:
- Gozal D. Sleep-disordered breathing and school performance in children. Pediatrics.1998;102 :616– 620
- Gozal D, Pope DW Jr. Snoring during early childhood and academic performance at ages thirteen to fourteen years. Pediatrics.2001;107 :1394– 1399
- ↵Marcus CL, Carroll JL. Obstructive sleep apnea syndrome. In: Loughlin GM, Eigen H, eds. Respiratory Disease in Children: Diagnosis and Management. 1st ed. Baltimore, MD: Williams & Wilkins, 1994:475–499
- ↵Nieminen P, Lopponen T, Tolonen U, Lanning P, Knip M, Lopponen H. Growth and biochemical markers of growth in children with snoring and obstructive sleep apnea. Pediatrics.2002;109(4) . Available at: www.pediatrics.org/cgi/content/full/109/4/e55
- ↵Reuveni H, Simon T, Tal A, Elhayany A, Tarasiuk A. Health care services utilization in children with obstructive sleep apnea syndrome. Pediatrics.2002;110 :68– 72
- Otake K, Delaive K, Walld R, Manfreda J, Kryger MH. Cardiovascular medication use in patients with undiagnosed obstructive sleep apnea. Thorax.2002;57 :417– 422
- ↵Kryger HM, Roos L, Delaive K, Walld R, Horrocks J. Utilization of health care services in patients with severe obstructive sleep apnea. Sleep.1996;19(suppl) :111– 116
- ↵Sanner BM, Klewer J, Trumm A, Randerath W, Kreuzer I, Zidek W. Long-term treatment with continuous positive airway pressure improves quality of life in obstructive sleep apnea syndrome. Eur Respir J.2000;16 :118– 122
- ↵Rechtschaffen A, Kales A. A Manual of Standardized Terminology: Techniques and Scoring System for Sleep Stages of Human Subjects. Washington, DC: US Public Health Service; 1968
- ↵World Health Organization. Collaborating Center for Drug Statistic Methodology Guidelines for ATC Classification and DDD Assignment. Oslo, Norway; WHO Publication Center; 2000
- ↵Thompson SG, Barber J. How should cost data in pragmatic randomized trials be analyzed? BMJ.2000;320 :1197– 1200
- ↵Marcus CL, Chapman D, Ward SD, et al. Clinical practice guideline: diagnosis and management of childhood obstructive sleep apnea syndrome. Pediatrics.2002;109 :704– 712
- Copyright © 2004 by the American Academy of Pediatrics