| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Sleep Disorders Center (Drs. Smith, Ronald, and Kryger, and Mr. Delaive), St. Boniface General Hospital Research Center; and the Center for Health Policy and Evaluation (Dr. Manfreda and Mr. Walld), Department of Community Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada.
Correspondence to: Meir H. Kryger, MD, FCCP, Sleep Disorders Center, St. Boniface General Hospital, R2034, 351 Tache Ave, Winnipeg, Manitoba, Canada R2H 2A6; e-mail: kryger{at}sleep umanitoba.ca
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
Objectives: To examine the causes of increased utilization, and what patients with OSAS were being treated for prior to this diagnosis.
Methods: We compared the records of 773 patients with OSAS to those of age-, gender-, geographic-, and physician-matched control subjects from the general population.
Results: We found that sleep apnea patients used 23 to 50% more resources (defined by physician fees, physician visits, and hospital nights) in the 5 years prior to diagnosis than did control subjects. We examined the diagnoses made and found that apnea patients are at higher risk for hypertension (odds ratio [OR], 2.5; 95% confidence interval [CI], 2.0 to 3.3), congestive heart failure (OR, 3.9; 95% CI, 1.7 to 8.9), cardiac arrhythmias (OR, 2.2; 95% CI, 1.2 to 4.0), cardiovascular disease (OR, 2.6; 95% CI, 2.0 to 3.3), chronic obstructive airways disease (OR, 1.6; 95% CI, 1.2 to 2.0), and depression (OR, 1.4; 95% CI, 1.0 to 1.9). To control for the confounding effects of obesity and to determine the independent effects of body mass index (BMI), gender, age, degree of hypoxemia, apnea-hypopnea index, and sleepiness in the 773 patients, we performed a logistic regression analysis with the dependent variable being diagnosis, and a linear regression analysis with the dependent variable being measures of health-care utilization. Age and BMI were significant independent predictors of most cardiovascular diagnoses and arthropathy. Male gender predicted ischemic heart disease (OR, 2.98; 95% CI, 1.36 to 6.54), and female gender was predictive of chronic obstructive airways disease (OR, 2.63; 95% CI, 1.85 to 3.72) and depression (OR, 2.24; 95% CI, 1.45 to 3.44). The best model predicting health-care utilization measures was comprised of age, gender, and BMI, and explained 9%, 14%, and 8% of the variability in physician fees, number of physician claims, and number of physician visits, respectively.
Conclusion: Of all comorbid diagnoses, significantly increased utilization was found for cardiovascular disease and especially hypertension in the OSAS patients.
Key Words: apnea • depression • health-care utilization • hypertension • medical economics • obstructive sleep apnea syndrome • sleep
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
OSAS, which is a common disorder affecting 2 to 4% of the adult general population, remains underdiagnosed and undertreated.3 4 This may be due in part to the commonly held belief by some that OSAS may not pose a serious health risk.5 Thus, the diagnosis and treatment of this condition may have a low priority in health-care systems.
At the time of diagnosis, patients often report a long history of symptoms going back many years. It has been shown that OSAS patients are heavy users of health-care resources, not only at the time of diagnosis, but also for years prior to diagnosis.6 7 It has also been found that diagnosis of OSAS and adherence to treatment results in a significant reduction in resource utilization (physician claims and hospital stays).8
The role of obesity in illness and its cost have been investigated9 10 ; however, these factors have not been studied in OSAS patients, who are often obese. There are no reports concerning the effect of body mass index (BMI), apnea severity, and gender on utilization of health-care resources in patients with OSAS. Our aim was to determine what patients are being treated for in the 5 years prior to the diagnosis of sleep apnea, what variables are associated with these diagnoses, and the use of health-care resources. We hypothesized that increased health-care utilization would be for a variety of diagnoses because of the broad spectrum of disorders and presentations associated with OSAS.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
Confidentiality of OSAS patients and control subjects was ensured by "encrypting" each person’s health insurance number and using the encrypted number as that person’s only unique identifier. This project was approved by the Human Ethics Committee of the University of Manitoba and the Access and Confidentiality Committee of Manitoba Health.
Selection of Patients
The data were collected at a university-based sleep-disorders
center. We selected all patients with polysomnographically proven OSAS
who had health-care utilization information in the MHdb going back 5
years before their laboratory diagnosis of sleep apnea.
Evaluation of Patients
All patients were Manitoba residents referred to the same
sleep-disorders center for assessment of OSAS. They were all evaluated
by one of the authors and underwent overnight polysomnography. This
involved multichannel monitoring, including EEG, electro-oculography,
electromyography, ECG, arterial oxygen saturation
(SaO2), and end-tidal carbon dioxide
levels. Thoracic and abdominal movements were continuously monitored by
belt plethysmography; activity in the anterior tibialis muscles was
monitored by standard electrodes. Diagnosis of OSAS was based on
history and polysomnographic findings.
Control Subjects
Using the MHdb, one control subject from the general population
was obtained for each OSAS patient. The control subjects were matched
to the OSAS patients for age, gender, postal code, and the most
frequent physician seen in an ambulatory setting in the previous 2
years. Matching for postal code was done to correct for socioeconomic
factors and distance to health-care services. Matching for a specific
primary physician was done to minimize biases that could occur if
patients and control subjects had different doctors. Seven hundred
sixty-six (99%) of the matches were exact for age. However, in the
event that an age match within 1 year could not be obtained, up to a
5-year age difference was allowed. Patients were not matched to control
subjects for BMI because that information is not included in the MHdb.
The authors were not permitted to contact the control subjects because
of privacy legislation.
Exclusion
Any case patients or control subjects with extreme health-care
usage (> 50 days in hospital over 5 years) or who were
institutionalized for chronic illness, or who required dialysis were
excluded from the study. This exclusion was done to limit our sample to
"typical" OSAS patients and "typical" control subjects and
resulted in exclusion of 31 of 809 eligible case patients. An
additional five case patients were excluded due to lack of an
appropriate matchable control subject. Our final working database
consisted of 773 OSAS patients and their matched control subjects.
Statistical Analysis
In addition to descriptive statistics, we used the t
test to compare continuous variables between groups (female vs male
subjects) and the paired t test for comparison between
matched case patients and control subjects. Wilcoxon’s test was used
when nonparametric procedure appeared appropriate. Nonconditional and
conditional logistic regression were used to calculate odds ratios
(ORs) with 95% confidence intervals (CIs). Linear trend was tested
using analysis of variance. Results were considered significant at
p < 0.05.12
We first compared case patients and their
matched control subjects. To determine the independent effects of BMI,
gender, age, degree of hypoxemia, apnea-hypopnea index (AHI), and
sleepiness in the 773 patients, we performed a logistic regression
analysis with the dependent variable being diagnosis, and a linear
regression analysis with the dependent variable being measures of
health-care utilization.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
|
|
|
|
Adjusted Predictors of Comorbid Diagnoses in OSAS Patients
To control for the nonmodifiable risk factors (age, gender),
weight (BMI), sleepiness (ESS), and sleep study findings (AHI,
percentage of time with SaO2
< 90%), we performed a logistic analysis with the dependent variable
being ICD-9 diagnosis in the group of 773 patients. The ESS is a
subjective measure of sleepiness, with values ranging from 0 (no
sleepiness) to 24 (sleepiness in most situations).13
14
The results are summarized in Table 4 .
Female patients with sleep apnea were more than twice as likely
as male patients to have chronic obstructive airways disease (OR, 2.63;
95% CI, 1.85 to 3.72) and depression (OR, 2.24; 95% CI, 1.45 to
3.44). Conversely, male sleep apnea patients were almost three times as
likely to have ischemic heart disease (OR, 2.98; 95% CI, 1.36 to 6.54)
as female patients. All cardiovascular diagnoses (hypertension,
ischemic heart disease, congestive heart failure, arrhythmia,
cardiovascular disease) and arthropathy were predicted by age. BMI was
a significant predictor of several diagnoses, including hypertension,
ischemic heart disease, congestive heart failure, cardiovascular
disease, and arthropathy. AHI, ESS, and percentage of time with
SaO2 < 90% added little to predict
comorbidity in OSAS patients.
|
To further delineate the role of obesity in health-care utilization of
sleep apnea patients, we stratified the 773 OSAS patients for whom we
had reliable BMI data based on the BMI groupings as defined by the
World Health Organization Report on Obesity15
and the
Expert Panel on the Identification, Evaluation, and Treatment of
Overweight and Obesity in Adults.16
The results are shown
in Table 5
. Linear trend analysis of the individual usage measures across the five
weight categories did not show a statistically significant trend for
any of the three measures (p = 0.927 for physician fees, p = 0.830
for number of physician claims, and p = 0.359 for number of physician
visits). Interestingly, the normal weight OSAS patients used resources
at similar rates to the most obese (BMI 
40
kg/m2) group for all three measures of
utilization. When the linear trends were reanalyzed for the overweight
groups (after removing the normal-weight sleep apnea patients), both
the number of physician claims (p = 0.031) and the number of
physician visits (p = 0.001) showed significantly increasing trends
with increasing obesity. Total physician fees (p = 0.089) did not
show a significant trend after reanalysis.
|
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
For five years before sleep apnea diagnosis, OSAS patients used medical resources at significantly higher rates than the control subjects. Analysis of three markers of usage, namely physician fees, number of physician visits, and number of nights spent in hospital, demonstrates the greater utilization in the OSAS patient group by 23 to 50% compared to the matched control subjects.
Previous work6 7 has shown an even greater difference in health-care utilization between OSAS patients and matched control subjects in the years prior to sleep apnea diagnosis than the current study. One of the major methodologic differences in this study is the matching of OSAS patients and control subjects for a specific primary physician. This was done to minimize biases in referral patterns. However, this may have selected out a group of control subjects who are more likely to seek regular physician contact and are sicker than the general population, and thus use more health-care resources. This only strengthens the association of OSAS with increased health-care usage prior to diagnosis and treatment.
Although we matched case patients and control subjects by age, gender, area of residence, and family doctor, we were precluded from matching by BMI because we were not permitted to contact the control subjects to obtain their BMI because of legislation protecting patient confidentiality. We, therefore, adjusted for the possible confound of weight, as well as gender, age, AHI, oxygen desaturation, and subjective sleepiness, by determining the independent effect of each of these variables (controlling for the others) on diagnosis as well as measures of health-care utilization within the patient group itself. We did not control for smoking and alcohol use. Peppard et al17 found no evidence that these were important confounders in linking apnea to hypertension. Nieto et al18 similarly reported that controlling for smoking and alcohol use had very little impact on their findings in examining the link between apnea and hypertension.
Because obesity is so often seen in OSAS, there is a tendency to focus on obesity per se as the cause of the increased morbidity (eg, hypertension) of these patients. Although overweight is a possible confounder between OSAS and morbidities such as hypertension, it has been suggested by Nieto et al18 that data linking OSAS to hypertension are also consistent with an alternate model, whereby sleep apnea is one of the mechanisms causing hypertension in the obese. Nieto et al18 thus suggest that adjusting for BMI may be an "overadjustment." When Peppard et al17 examined the relationship between AHI and hypertension, adjusting for more factors than age and gender did not change any of the findings from statistically significant to nonsignificant. Indeed, when we dealt with the possible confounding variables in our OSAS patients, we found that the impact of BMI was quite modest. People who are markedly overweight are not necessarily heavy users of health-care resources. We recently studied the data of a group of "healthy" obese individuals (mean BMI, 43.4 kg/m2) who participated in a Canadian population-based study of cardiovascular risk factors. They had health-care utilization over a 7-year period that was quite similar to a group of individuals chosen from the general population but matched for age, gender, and postal code.19
The best model predicting measures of health-care utilization in OSAS was comprised of age, gender, and BMI; it explained < 15% of the variability in any individual utilization measure. We then further studied the role of BMI by category in predicting usage and found that even normal-weight OSAS patients have extreme usage. Linearity for two of the three usage measures (number of physician claims and number of physician visits) was shown for the overweight and obese groups; however, when all weight groups were considered in the analysis, none of the trends in usage measures were statistically significant. This may be somewhat surprising at first glance. However, as is the experience in most sleep laboratories, thin patients may have severe sleep apnea (caused, for example, by retrognathia), and very obese patients may not have apnea at all or may have mild apnea. Even within the obese population there is no a priori reason why, for example, a patient with a BMI of 38 kg/m2 should have greater expenditure than a patient with a BMI of 34 kg/m2. Similarly, the lack of correlation with AHI is of interest. Again, this is due to the fact that AHI is probably an imperfect linear measure of severity. For example, 30 long apneic episodes per hour with severe oxygen desaturation may have greater physiologic impact than 80 short episodes with little hypoxemia. This highlights the fact that there is no one measure that by itself adequately defines the severity of the disorder and would therefore predict health-care utilization.
When we looked at diagnosis specific expenditure, we found
significantly increased usage in the OSAS group for hypertension and
cardiovascular disease. This is not surprising, in that
others20
have recently shown OSAS to be an independent
risk factor for the development of hypertension. In a large
community-based cross-sectional analysis from the Sleep Heart Health
Study, Nieto et al18
showed ORs for hypertension (adjusted
for demographics, anthropometric variables, alcohol intake, and
smoking) of 1.37 (95% CI, 1.03 to 1.83) and 1.46 (95% CI, 1.12 to
1.88) in comparing the highest and lowest categories of AHI ( 30
events per hour vs < 1.5 events per hour) and percentage sleep time
at < 90% oxygen saturation ( 12% vs < 0.05%), respectively.
In a prospective population-based study, Peppard and
colleagues17
showed ORs for hypertension (adjusted for
base-line hypertension status, age, sex, habitus variables, and alcohol
and cigarette use) ranging from 1.42 to 2.89 at 4-year follow-up based
on baseline AHIs.
The health-care utilization costs that we have determined actually underestimate expenditure because they do not include the costs of medications. To estimate the medication costs, we examined a new database that includes medication use by all residents of Manitoba. This database allowed us to examine drug use in a subset of 422 OSAS patients (we were unable to examine the data in the patients who received a diagnosis prior to the establishment of this database), and we found that OSAS patients were much more likely to receive antihypertensive medications (OR, 2.5; 95% CI, 1.3 to 4.8) than control subjects. The average hypertensive patient with sleep apnea received more prescriptions during the course of a year (6.4 prescriptions vs 4.4 prescriptions) than hypertensive control subjects, and the average prescription was more expensive ($39.88 vs $33.92), resulting in higher annual medication use. Thus, it appears that treatment of hypertension in an OSAS patient is more expensive than treatment of hypertension in an individual from the general population. This is consistent with the previously published observation that patients with sleep apnea may have difficult-to-treat hypertension.21 Future studies will examine medication use in detail.
Although depression was more commonly diagnosed in the OSAS patients than in the control subjects, a paradoxical finding was that medical costs associated with depression in the control subjects were actually higher than those in OSAS patients. On examining the services associated with this increased expenditure, we found that psychotherapy was much more commonly used in the treatment of the control subjects with depression compared to OSAS patients with depression. We hypothesized that sleep apnea patients who are sleepy would not be good candidates for such therapy and might be more likely to be prescribed medication to treat their depression. This was confirmed when we analyzed the annual use of antidepressants. We found that the average sleep apnea patient receiving antidepressants used $324.76 of medication compared to $232.21 in the control subjects receiving antidepressants. The average prescription in the OSAS patients was $60.17 vs $38.97 for the control subjects. This supports the hypothesis that sleep apnea patients with a diagnosis of depression are treated differently than control subjects with depression, and that they are more likely to receive drug therapy. What might be even more disturbing is the possibility that sleep apnea patients are being treated for an illness (depression) that they may not have. This is because some of the classic symptoms of apnea (sleepiness, loss of energy) may be misinterpreted to represent depression.
|
Conclusion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
|
Footnotes |
|---|
Supported in part by National Institutes of Health grant No. R01 HL63342–01A1.
Received for publication April 9, 2001. Accepted for publication July 24, 2001.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. R. Chasens and E. Olshansky The Experience of Being Sleepy While Managing Type 2 Diabetes Journal of the American Psychiatric Nurses Association, October 1, 2006; 12(5): 272 - 278. [Abstract] [PDF]
|
|
|
|
 |
|
 A. Tarasiuk, S. Greenberg-Dotan, T. Simon, A. Tal, A. Oksenberg, and H. Reuveni Low socioeconomic status is a risk factor for cardiovascular disease among adult obstructive sleep apnea syndrome patients requiring treatment. Chest, September 1, 2006; 130(3): 766 - 773. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Insalaco, S. Romano, O. Marrone, A. Salvaggio, and G. Bonsignore A New Method of Negative Expiratory Pressure Test Analysis Detecting Upper Airway Flow Limitation To Reveal Obstructive Sleep Apnea Chest, October 1, 2005; 128(4): 2159 - 2165. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Tarasiuk, S. Greenberg-Dotan, Y. S. Brin, T. Simon, A. Tal, and H. Reuveni Determinants Affecting Health-Care Utilization in Obstructive Sleep Apnea Syndrome Patients Chest, September 1, 2005; 128(3): 1310 - 1314. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D Schlosshan and M W Elliott Sleep * 3: Clinical presentation and diagnosis of the obstructive sleep apnoea hypopnoea syndrome Thorax, April 1, 2004; 59(4): 347 - 352. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Tarasiuk, T. Simon, A. Tal, and H. Reuveni Adenotonsillectomy in Children With Obstructive Sleep Apnea Syndrome Reduces Health Care Utilization Pediatrics, February 1, 2004; 113(2): 351 - 356. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. C. Netzer, J. J. Hoegel, D. Loube, C. M. Netzer, B. Hay, R. Alvarez-Sala, and K. P. Strohl Prevalence of Symptoms and Risk of Sleep Apnea in Primary Care Chest, October 1, 2003; 124(4): 1406 - 1414. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I.A. Harsch, P.C. Konturek, C. Koebnick, P.P. Kuehnlein, F.S. Fuchs, S. Pour Schahin, G.H. Wiest, E.G. Hahn, T. Lohmann, and J.H. Ficker Leptin and ghrelin levels in patients with obstructive sleep apnoea: effect of CPAP treatment Eur. Respir. J., August 1, 2003; 22(2): 251 - 257. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Stores Misdiagnosing sleep disorders as primary psychiatric conditions Advan. Psychiatr. Treat., January 1, 2003; 9(1): 69 - 77. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K Otake, K Delaive, R Walld, J Manfreda, and M H Kryger Cardiovascular medication use in patients with undiagnosed obstructive sleep apnoea Thorax, May 1, 2002; 57(5): 417 - 422. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
