Abstract
Obstructive sleep apnea may lead to complications if not identified and treated. Polysomnography is the diagnostic standard, but is often inaccessible due to bed shortages. A system that facilitates prioritization of patients requiring sleep studies would thus be useful. We retrospectively compared the accuracy of a two-stage risk-stratification algorithm for sleep apnea using questionnaire plus nocturnal pulse oximetry against using polysomnography to identify patients without apnea (Objective 1) and those with severe apnea (Objective 2). Patients were those referred to a university-based sleep disorders clinic due to suspicion of sleep apnea. Subjects completed a sleep apnea symptom questionnaire, and underwent oximetry and two-night polysomnography. We used bootstrap methodology to maximize sensitivity of our model for Objective 1 and specificity for Objective 2. We calculated sensitivity, specificity, positive and negative predictive values, and rate of misclassification error of the two-stage risk-stratification algorithm for each of our two objectives. The model identified cases of sleep apnea with 95% sensitivity and severe apnea with 97% specificity. It excluded only 8% of patients from sleep studies, but prioritized up to 23% of subjects to receive in-laboratory studies. Among sleep disorders clinic referrals, a two-stage risk-stratification algorithm using questionnaire and nocturnal pulse oximetry excluded few patients from sleep studies, but identified a larger proportion of patients who should receive early testing because of their likelihood of having severe disease.
Keywords: obstructive sleep apnea; risk stratification; polysomnography; nocturnal pulse oximetry; questionnaire
Identifying obstructive sleep apnea (OSA) syndrome is important because it leads to adverse outcomes if left untreated (1-4), and because it improves with treatment (5). Although in-laboratory polysomnography (PSG) is an accurate diagnostic tool, it requires significant expertise and expense (6) and is often associated with long waiting periods (7).
The use of PSG in a sleep clinic population would be optimized if a risk stratification strategy were developed that achieves the following two goals. First, it identifies patients whose risk for sleep apnea is so low that PSG is not warranted. Second, it identifies the subset of patients who are at risk for severe OSA (8), who should be prioritized for early testing. To fulfill these two objectives, as well as to maintain accuracy while containing cost compared with gold standard tests, we developed a model that relies on a two-stage method of identifying cases. We based our model on existing prototypes for other conditions (9-15). The two component tests that we chose for our algorithm were the Multivariable Apnea Prediction (MAP) questionnaire (16) and nocturnal pulse oximetry (nPO) (5, 17-20). The objectives of this two-stage algorithm are to avoid doing sleep studies on patients at low risk for sleep apnea, and to prioritize scheduling of patients who are likely to have severe disease. We tested the hypothesis that our strategy could meet these two objectives in a sleep center population.
We selected subjects from patients presenting to our sleep center for evaluation for possible sleep apnea. All completed the MAP questionnaire (16), plus a full-night sleep study (PSG) with concurrent oximetry. We excluded subjects who were already diagnosed with OSA or obesity-hypoventilation syndrome or used supplemental oxygen.
Sleep studies were completed and scored by polysomnographic technologists at our sleep center according to the method of Rechtschaffen and Kales (21). Twelve-channel recordings were done on all patients that included monitoring of the electroencephalogram, eye movements, tibialis anterior and chin electromyography, respiratory effort, snoring, airflow, and oximetry. The American Sleep Disorders Association (ASDA) definition was used to score arousals during sleep (22). The respiratory disturbance or apnea–hypopnea index (RDI, AHI) was defined as the number of apneas plus hypopneas divided by the total sleep time in hours. An apnea was defined as complete cessation of airflow for at least 10 s and a hypopnea was defined
by a ⩾ 50% reduction in airflow for at least 10 s, associated with a ⩾ 4% fall in oxyhemoglobin saturation or an arousal. We judged OSA to be present if the RDI was ⩾ 5 events/h, and severe OSA to be present if the RDI was ⩾ 30 events/h (23).
As described previously (16), we determined each subject's MAP score, which predicts apnea risk using a score between 0 and 1, with 0 representing low risk and 1 representing high risk. A single observer then scored the oximetry strips without knowledge of the PSG results. The oximetry desaturation index (ODI) was the number of desaturations divided by the total test time. We obtained desaturation indices using both 3% and 4% drops, that is, ODI3 and ODI4, respectively. On PSG, we judged OSA to be present if the respiratory disturbance index (RDI) was ⩾ 5 events/h (case 1) and severe OSA to be present if the RDI was ⩾ 30 events/h (case 2) (23).
Using ROCKIT software (24), we performed “area under the curve” (AUC) analysis for receiver operating characteristic (ROC) curves (25) to determine relative discriminatory power of the MAP, ODI3, and ODI4 (26) and to determine their sensitivity and specificity (27).
We applied our algorithm retrospectively on all included subjects. We used a predefined upper bound (UB) value and lower bound (LB) value for the MAP score to separate the subjects into three groups. Those who had high MAP scores (MAP > UB) were predicted to have OSA and had subsequent review of PSG to see if this were so. Those with low MAP scores (MAP < LB) were predicted to be free of OSA and would not be further studied. Those with MAP scores between UB and LB underwent scoring of nocturnal pulse oximetry, and the ODI was compared against a predefined threshold (ODIthreshold) desaturation index. Those with ODI ⩾ ODIthreshold were predicted to have OSA and hence would undergo PSG evaluation, whereas those with ODI < ODIthreshold were predicted to be free of OSA.
Using Statistical Application Software (SAS) programming (Cary, NC), we computed our algorithm's sensitivity and specificity against PSG for each of 125 combinations of UB, LB, and ODIthreshold (i.e., 125 “parameter sets”) for case 1 and for case 2 separately (23).
We first selected a random sample of 80% of the clinic population, the “estimation sample.” We generated 200 bootstrap resamples from this estimation sample. To the 125 possible parameter sets associated with each bootstrap resample, we applied our criterion function to select the optimum parameter set (or sets). We then averaged all the optima from the 200 resamples to find what we called the unbiased estimate of the optimum parameter set. We did this for case 1 and for case 2 separately.
We applied our algorithm, using the two unbiased optimal parameter sets, to our reserved “validation sample” of 20% of our original clinic population. We computed sensitivity, specificity, and positive and negative predictive values of our two models separately, that is, for predicting RDI ⩾ 5/h and RDI ⩾ 30/h.
Of 421 patients considered for the study, 359 met the inclusion criteria. We excluded 62 subjects—46 because they had been ordered to have “split-night” studies by their referring physician, 12 because they were previously diagnosed with OSA, 3 because they were already receiving supplemental oxygen, and 1 because of an incomplete MAP questionnaire. The final group consisted of 243 (67.7%) men and 116 (32.3%) women. Mean age ± SD was 47.2 ± 13.2 yr. The racial composition was 72.2% white, 24.2% black, and 3.6% other. The average body mass index (BMI) ± SD was 32.4 ± 8.8 kg/m2. The MAP frequency distribution (Figure 1) shows a high prevalence of clinic patients with MAP scores exceeding 0.6. The mean ± SD of the RDI scores was 25.9 ± 29.5 events/h, with median 12 events/h, and range 0 to 110 events/h. Using RDI ⩾ 5/h, ⩾ 15/h, and ⩾ 30/h as the diagnostic criterion, the prevalence of OSA in this sleep center population was 69.4%, 47.1%, and 33%, respectively.
Fig. 1. Interval-specific MAP frequencies of sleep center population in each of the 10 MAP deciles. The distribution is right skewed, with a heavy proportion of the sleep clinic population having MAP scores exceeding a value of 0.6.
[More] [Minimize]We generated a receiver operating characteristic curve using MAP values against RDI ⩾ 5/h (Figure 2, upper panel ) and RDI ⩾ 30/h (Figure 2, lower panel ). For RDI ⩾ 5/h, the respective values of AUC (± SE) for MAP, ODI3, and ODI4 were 0.834 (± 0.024), 0.947 (± 0.011), and 0.936 (± 0.012). For RDI ⩾ 30/h, AUC for ODI4 was 0.98 (± 0.008), showing slightly improved discrimination of oximetry when using more stringent criteria for defining the presence of sleep apnea. For ODI4, sensitivity and specificity were 85% and 90%, respectively, using RDI ⩾ 5/h, and 90% and 95% using RDI ⩾ 30/h. Similar values were obtained for ODI3 (see Figure 2 and Table 1).
Fig. 2. Receiver operating characteristic curves using MAP questionnaire (thin solid line) or oximetry (heavy solid line). To score oximetry, we used ODI3, the number of episodes of fall in saturation ⩾ 3%/h of test time. RDI ⩾ 5/h (upper panel ) or RDI ⩾ 30/h (lower panel ) defined OSA. Compared with reference (straight dashed line), MAP shows good discriminatory power, and ODI3 better. The curve for ODI4, which used ⩾ 4% fall in saturation/h of test time, was superimposed on that for ODI3 and is not displayed separately.
[More] [Minimize]| AUC | Sensitivity | Specificity | ||||
|---|---|---|---|---|---|---|
| Discriminatory power using RDI ⩾ 5 events/h | ||||||
| MAP | 0.834 | 0.819 | 0.700 | |||
| ODI3 | 0.947 | 0.860 | 0.909 | |||
| ODI4 | 0.936 | 0.852 | 0.900 | |||
| Reference | 0.500 | 1.000 | 1.000 | |||
| Discriminatory power using RDI ⩾30 events/h | ||||||
| MAP | 0.786 | 0.805 | 0.631 | |||
| ODI3 | 0.981 | 0.908 | 0.942 | |||
| ODI4 | 0.979 | 0.899 | 0.946 | |||
| Reference | 0.500 | 1.000 | 1.000 | |||
We computed the sensitivity and specificity of our algorithm as a function of the value we chose for UB, LB, and ODIthreshold. We varied UB from 0.5 to 0.9 and LB from 0.1 to 0.5 in 0.1 unit increments, and ODIthreshold from 5 to 25 events/h by increments of 5. We computed sensitivity and specificity for these 125 combinations of UB, LB, and ODIthreshold.
As Table 2 shows, for a given ODIthreshold (15/h), if UB is kept constant at a specific value (0.7), varying LB has negligible effect on either sensitivity or specificity of the algorithm compared with PSG for identifying patients with RDI ⩾ 5 events/h (left two columns). This effect is also true for RDI ⩾ 30 events/h (right two columns). This effect is true for all levels of ODIthreshold and for any value of MAP UB that we studied (data not shown).
| Upper Bound = 0.7 | ||||||||
|---|---|---|---|---|---|---|---|---|
| Lower Bound | RDI ⩾ 5/h ODIthreshold = 15/h | RDI ⩾ 30/h ODIthreshold = 15/h | ||||||
| Sensitivity | Specificity | Sensitivity | Specificity | |||||
| 0.1 | 0.724 | 0.836 | 0.958 | 0.647 | ||||
| 0.2 | 0.720 | 0.836 | 0.958 | 0.651 | ||||
| 0.3 | 0.716 | 0.836 | 0.950 | 0.651 | ||||
| 0.4 | 0.716 | 0.845 | 0.950 | 0.656 | ||||
| 0.5 | 0.696 | 0.845 | 0.916 | 0.660 | ||||
The effect of varying UB is less straightforward, but still consistent and predictable (Figures 3 and 4). The effect of UB interacts with the effect of ODIthreshold. First, we discuss the case when RDI ⩾ 5/h is used to define apnea. UB has a negligible effect on sensitivity when ODIthreshold is low. As an example, for ODIthreshold = 5/h, LB = 0.3, as UB is increased from 0.5 to 0.9, the sensitivity drops very little (Figure 3, upper panel, dashed black line). At larger values of ODIthreshold, however, the effect of raising UB on sensitivity increases considerably. For example, at ODIthreshold = 25 events/h, increasing UB from 0.5 to 0.9 while LB = 0.3 results in dramatic decline in sensitivity (Figure 3, lower panel, solid black line) and dramatic increase in specificity (Figure 3, lower panel, solid gray line). Moreover, the effect of varying UB on sensitivity is less pronounced than its effect on specificity for any value of ODIthreshold. An interaction between ODIthreshold and LB was not observed (data not shown). Table 3 (left four columns) provides a numerical synopsis of these data.
Fig. 3. Upper bound–ODIthreshold interaction in predicting RDI ⩾ 5 events/h for a representative lower bound value of 0.3. Specificity (lower panel ) and particularly sensitivity (upper panel ) change more prominently with upper bound when ODIthreshold is 25 events/h (solid lines), compared with the lower value of 5 events/h (dashed lines).
[More] [Minimize]
Fig. 4. Upper bound–ODIthreshold interaction in predicting RDI ⩾ 30 events/h for a representative lower bound value of 0.3. As upper bound is increased from 0.5 to 0.9 (x-axis), we see greater changes in sensitivity (upper panel ) and specificity (lower panel ) when ODIthreshold is high at 25/h (solid lines), compared with when ODIthreshold is low, at 5/h (dashed lines).
[More] [Minimize]| Upper Bound | Lower Bound = 0.3 | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RDI ⩾ 5/h ODIthreshold | RDI ⩾ 30/h ODIthreshold | |||||||||||||||
| 5/h | 25/h | 5/h | 25/h | |||||||||||||
| Sensitivity | Specificity | Sensitivity | Specificity | Sensitivity | Specificity | Sensitivity | Specificity | |||||||||
| 0.5 | 0.932 | 0.600 | 0.880 | 0.627 | 0.983 | 0.336 | 0.966 | 0.394 | ||||||||
| 0.6 | 0.916 | 0.682 | 0.796 | 0.709 | 0.983 | 0.390 | 0.958 | 0.515 | ||||||||
| 0.7 | 0.876 | 0.800 | 0.672 | 0.855 | 0.975 | 0.481 | 0.899 | 0.680 | ||||||||
| 0.8 | 0.840 | 0.873 | 0.512 | 0.936 | 0.975 | 0.552 | 0.832 | 0.851 | ||||||||
| 0.9 | 0.816 | 0.909 | 0.400 | 0.982 | 0.975 | 0.593 | 0.790 | 0.967 | ||||||||
When RDI ⩾ 30/h is used to define apnea (Figure 4), the interaction between UB and ODIthreshold was still present. Upper bound's effect on sensitivity (upper panel) and specificity (lower panel) was more pronounced at the higher ODIthreshold value of 25/h rather than the lower value of 5/h (compare solid lines to dashed lines).
The comparative results of these two definitions of OSA (⩾ 5/h and ⩾ 30/h) are shown numerically in Table 3.
As Figure 5 (white bars) shows, apnea prevalence (RDI ⩾ 5/h) increases with MAP score, ranging from 0% (for MAP 0.0– 0.1) to 93% (for MAP 0.8–0.9). The overall prevalence of OSA in the population as a whole was 69%. The prevalence of severe apnea (RD ⩾ 30 events/h, solid bars) increased with MAP score. For those with MAP ⩾ 0.4, prevalence of severe apnea ranged from 12% to 71%, while only one patient with an MAP score below 0.4 had severe OSA.
Fig. 5. Prevalence of OSA by MAP interval in clinic population. OSA prevalence is high in nearly all intervals in the sleep center population, and particularly so in those with MAP ⩾ 0.6. This was true regardless of the cutpoint for OSA: RDI ⩾ 5 events/h or 30 events/h.
[More] [Minimize]Using our criterion functions for each of our two objectives, we obtained the biased optimal parameter set. For Objective 1, this set was UB = 0.58 (range 0.5–0.8), LB = 0.14 (range 0.1–0.2), and ODIthreshold = 5.02 desaturations/h (range 5–10). For Objective 2, that is, to detect those with RDI ⩾ 30 events/h who would be suitable for split-night studies, the biased optimal parameter set was UB = 0.9 (range 0.9–0.9), LB = 0.38 (range 0.2–0.5), and ODIthreshold = 21/h (range 10–25). Interestingly, the entire bootstrap distribution of UB values was fixed at 0.9. The results are summarized in Table 4. Table 4 also shows the total number of parameter sets (which resulted from the 200 bootstrap resamples), which were averaged to obtain the biased optimal parameter set for each of our two objectives.
| Parameter | Bootstrap Mean* | Min* | Max* | Number of Samples† | ||||
|---|---|---|---|---|---|---|---|---|
| Objective 1: Selection of Subjects for In-lab Sleep Study | ||||||||
| UB | 0.58 | 0.50 | 0.80 | 402 | ||||
| LB | 0.14 | 0.10 | 0.20 | 402 | ||||
| ODI | 5.02 | 5.00 | 10.00 | 402 | ||||
| Objective 2: Selection of Subjects for Split-night Testing | ||||||||
| UB | 0.90 | 0.90 | 0.90 | 237 | ||||
| LB | 0.38 | 0.20 | 0.50 | 237 | ||||
| ODI | 21.14 | 10.00 | 25.00 | 237 | ||||
Using our unbiased estimates of each of the two optimal parameter sets, we computed model accuracy in our validation sample, which consisted of 75 patients (20%) from the original
clinic population (Table 5).
| Estimation Sample* | 2.5% LCL | 97.5% UCL | Validation Sample† | |||||
|---|---|---|---|---|---|---|---|---|
| Objective 1: Selection of Subjects for In-lab Sleep Study | ||||||||
| Sensitivity | 0.948 | (0.913 | 0.975) | 0.941 | ||||
| Specificity | 0.684 | (0.575 | 0.806) | 0.667 | ||||
| PPV | 0.861 | (0.810 | 0.919) | 0.857 | ||||
| NPV | 0.865 | (0.790 | 0.935) | 0.842 | ||||
| Objective 2: Selection of Subjects for Split-night Testing | ||||||||
| Sensitivity | 0.845 | (0.769 | 0.925) | 0.833 | ||||
| Specificity | 0.967 | (0.938 | 0.989) | 0.947 | ||||
| PPV | 0.935 | (0.880 | 0.977) | 0.833 | ||||
| NPV | 0.920 | (0.879 | 0.958) | 0.947 | ||||
For Objective I, identifying all potential cases of apnea, the model shows a sensitivity of 95%, with a negative predictive value of 87%. The specificity and positive predictive values are 68% and 86%, respectively. These values are given in Table 5 (top), along with their 2.5% lower confidence limit and 97.5% upper confidence limits. The misclassification rate, defined as the sum of the total false positive and false negative rates divided by the total sample size (n = 75), was 17%. The false negative rate was very low at 7.9%, and the false positive rate was 9.2%.
For Objective II, identifying patients with severe apnea (RDI ⩾ 30/h) for split-night studies, we found that the model had a specificity of 97% and a positive predictive value of 94%. The sensitivity was 85%, and the negative predictive value was 92%. These values are also given in Table 5 (bottom). The false negative and false positive rates were 5.3% and 3.9%, respectively.
By comparison, the 20% validation sample yielded estimates of accuracy that were comparable to those obtained for the 80% estimation sample for both objectives. (Compare the first and fourth columns of Table 5.)
The prevalence of OSA (RDI ⩾ 5/h) in the estimation, validation, and total sample was 70%, 68%, and 69%, respectively. Using RDI ⩾ 30/h, these respective values were 35%, 24%, and 33%. Thus the three groups had similar prevalences of apnea and severe apnea.
Our first objective was to identify subjects who do not need a sleep study, which our model accomplished with 95% sensitivity. We reviewed the five cases of apnea that were missed by our algorithm, out of a total of 51 patients with apnea, in the model validation population (n = 75). The individual RDI scores of these five patients were 5, 5, 6, 10, and 11 events/h. Thus, all missed cases had mild disease.
The number of such missed cases is a function of the parameters we chose. Our specific parameter selection criteria resulted in an optimal LB value of 0.1. However, review of Figure 5 shows that most cases of OSA are associated with MAP ⩾ 0.4. If we use LB = 0.4 as the optimal parameter, we exclude 73/359 = 20% of subjects from unneeded sleep studies, and if we use LB = 0.1, we exclude only 25/359 = 8% of subjects, none of whom had apnea. Excluding this additional 12% of subjects by using LB = 0.4 instead of 0.1 comes at the cost of missing 13 additional cases of apnea, with RDI values 7–23/h for all but one, who had RDI 73/h. Thus, most patients who are missed when we use LB = 0.4 still have mild to moderate apnea (30). However, missing these additional cases might prove to be expensive in the long run, as evidence suggests that patients who have even mild OSA may be at risk for adverse effects, including hypertension (31, 32) and vehicular crashes (33), and may also benefit from therapy (34).
These results suggest that although our risk stratification algorithm works well, it does not remove the need for sleep studies in many patients. This is hardly surprising, as patients being evaluated had come with the clinical suspicion of sleep apnea, and the prevalence of the disorder in this sample is high (69%). This particular aspect of our algorithm might therefore have more utility if applied as a screening tool in a more general population. This is an area for future study. Importantly, this finding highlights the heterogeneity of clinical presentation among those with milder forms of sleep apnea, wherein subjects with similar disease severity may present with quite variable symptoms.
Identification of severe cases of OSA was our second objective, and our specificity was 97%. We correctly identified 15 of the 18 patients in the validation sample with RDI ⩾ 30 events/h, and 54 of the 57 patients with RDI < 30 events/h. We reviewed the RDI scores of patients incorrectly predicted to have severe apnea by our algorithm. Only three patients were incorrectly identified as having severe apnea, and these had RDI scores of 1, 20, and 25 events/h. Thus, two of three of these subjects had at least moderate apnea, characterized by RDI between 15 and 30 events/h. In addition, three cases of severe apnea were not prioritized by the model to undergo PSG, and these had RDI values of 34/h, 42/h, and 66/h. However, they would have been selected for PSG when the first model is applied, as they had MAP scores of 0.67, 0.80, and 0.71, respectively, all in the high MAP group for the first model. Hence, the diagnosis would not have been missed in any of them.
Thus, the risk stratification algorithm is more useful for prioritizing patients to receive sleep studies than for excluding apnea. Given its value, it is reasonable to consider whether we could select the range of MAP values that is critical. In our model, we found that patients who have an MAP ⩾ 0.9 should have sleep studies, independent of oximetry results. In the clinic population as a whole, 13 (4%) of 359 patients had MAP ⩾ 0.9, and 8 of these (62%) had RDI ⩾ 30/h. In addition, 11 of the 13 (85%) had RDI ⩾ 5/h. If we instead used MAP ⩾ 0.8, we would find that 82 of 359, or 23%, had MAP ⩾ 0.8, of whom 47 of 82 (57%) had RDI ⩾ 30/h, and 75 of 82 (91%) had RDI ⩾ 5/h. This is a substantial number of subjects. Thus, we believe that it is reasonable to propose that the sleep center use the optimal model parameters outlined here, or alternatively send the larger number of patients with MAP ⩾ 0.8 for full sleep studies. The latter represents a slightly adjusted algorithm, but could be used if overnight oximetry was not available, and moreover would remove the costs for this additional test.
Although these diagnostic rates are reasonably accurate, techniques other than MAP and oximetry may yield even better results, including other questionnaires (16, 35-40), relative proportions of intraoral measurements (41), or other simpler systems to monitor for respiratory disturbances that include readings of nasal pressure (42).
In interpreting these results, we note that oximetry was conducted concurrently with PSG, raising the concern that the two tests could not be scored independently. However, a second observer who had no knowledge of the PSG results scored the oximetry records. Additionally, random rescoring of nPO tracings by a second as well as by the original interpreter showed no significant differences in scores. Test–retest and interrater reliabilities (TRT and IRR) (43) measured using intraclass correlation coefficients were high, and were 0.997 and 0.981, respectively. Performing both nPO and PSG together had benefits, such as ensuring that the individual was sleeping in an identical position and was in an identical stage of sleep for both studies. Moreover, a technician was present to correct signal errors immediately. Future investigations should assess the predictive value of unattended oximetry conducted in the patient's home.
In our sensitivity analysis of oximetry, we observed the trends in sensitivity and specificity when we used a 3% drop in saturation instead of 4% to define events. We found negligible differences in sensitivity and specificity trends using this alternative method of interpreting oximetry data (data not shown).
Another strength of our study was that the prevalence of apnea was similar in the validation sample, in the estimation sample, and in the total sample, for either apnea definition— RDI ⩾ 5/h or RDI ⩾ 30/h. This suggests that the validation sample was representative of both the estimation sample and of the total sample.
In our algorithm, we explored the value of questionnaire and oximetry, but other techniques may yield even better results, including other questionnaires (16, 35-40). Relative proportions of intraoral measurements combined with BMI (41) have yielded an AUC of 0.996, with BMI having an exceptionally strong predictive value in this study, having an AUC of 0.938, compared with the value of 0.734 found by us previously (16). Other simpler systems to monitor for respiratory disturbances include readings of nasal pressure (42), and yielded a sensitivity of 100%, specificity of 92%, positive predictive value of 92%, and negative predictive value of 100%, as compared with PSG. These values were better than those of nPO, which produced sensitivity of 75%, specificity of 85%, positive predictive value of 75%, and negative predictive value of 89%. An approach that combines both nasal pressure measurement and oximetry may therefore be the optimal one for the second stage of our model.
In conclusion, our algorithm offers simplicity, feasibility, and sufficient accuracy. The model can omit unnecessary sleep studies in subjects who are unlikely to have OSA, and prioritize high-risk patients to receive early testing. The former reduces the number of sleep studies by only 8–12%, given the high prevalence of apnea in a sleep center population. It is likely to have greater utility in more general populations. The second objective addresses a larger percentage of the population (up to 23%), and gives sleep specialists a tool to identify high-risk patients who should be studied quickly. The optimal tests to be used in this two-stage method have yet to be determined. As diagnostic technology develops, we anticipate that this algorithm can be further refined.
This work was supported by NIH Grants HL-42236, HL-07713, and CRC-RR-00040.
| 1. | Flemons W, Tsai WQuality of life consequences of sleep-disordered breathing. J Allergy Clin Immunol991997S750S756 |
| 2. | Olson L, King M, Hensley M, Saunders NA community study of snoring and sleep-disordered breathing. Health outcomes. Am J Respir Crit Care Med1521995717720 |
| 3. | Phillipson ESleep apnea—a major public health problem. N Engl J Med328199312711273 |
| 4. | Shepard JJHypertension, cardiac arrhythmias, myocardial infarction, and stroke in relation to obstructive sleep apnea. Clin Chest Med131992437458 |
| 5. | Baumel M, Maislin G, Pack APopulation and occupational screening for obstructive sleep apnea: are we there yet? Am J Respir Crit Care Med1551997914 |
| 6. | Pack AObstructive sleep apnea. Adv Intern Med391994517567 |
| 7. | Pouliot Z, Peters M, Neufeld H, Kryger MUsing self-reported questionnaire data to prioritize OSA patients for polysomnography. Sleep201997232236 |
| 8. | Recommendations of criteria for measurements, definitions and severity ratings of sleep related breathing disorders in adults. Report of an American Sleep Disorders Association Task Force. Chicago, 1998. |
| 9. | Hochberg M, Schmidt M, Funk K, Sutton J, Stevens MValidity and yield of a two-stage screening procedure for systemic lupus erythematosus. Clin Exp Rheumatol119836771 |
| 10. | Martin D, Blackburn N, O'Connell K, Brant E, Goetsch EEvaluation of the World Health Organisation antibody-testing strategy for the individual patient diagnosis of HIV infection (strategy III). S Afr Med J851995877880 |
| 11. | Allen M, Embry BBiopsy of the prostate guided by transrectal ultrasonography: early experience in a teaching community hospital. S Med J841991579586 |
| 12. | Winawer S, Sherlock PSurveillance for colorectal cancer in average-risk patients, familial high-risk groups, and patients with adenomas. Cancer50198226092614 |
| 13. | Zavertnik J, McCoy C, Robinson D, Love NCost-effective management of breast cancer. Cancer69199219791984 |
| 14. | Bower S, Bewley S, Campbell SImproved prediction of preeclampsia by two-stage screening of uterine arteries using the early diastolic notch and color Doppler imaging. Obstet Gynecol8219937883 |
| 15. | Sen B, Wilkinson G, Mari JPsychiatric morbidity in primary health care. A two-stage screening procedure in developing countries: choice of instruments and cost-effectiveness. Br J Psychol15119873338 |
| 16. | Maislin G, Pack A, Kribbs N, Smith P, Schwartz A, Kline L, Schwab R, Dinges DA survey screen for prediction of apnea. Sleep181995158166 |
| 17. | Series F, Marc I, Cormier Y, La Forge JUtility of nocturnal home oximetry for case finding in patients with suspected sleep apnea hypopnea syndrome. Ann Intern Med1191993449453 |
| 18. | Yamashiro Y, Kryger MNocturnal oximetry: is it a screening tool for sleep disorders? Sleep181995167171 |
| 19. | Levy P, Pepin J, Deschaux-Blanc C, Paramelle B, Brambilla CAccuracy of oximetry for detection of respiratory disturbances in sleep apnea syndrome. Chest1091996395399 |
| 20. | Lacassagne L, Didier A, Murris-Espin M, Charlet J, Chollet P, Leophonte-Domairon M, Tiberge M, Pessey J, Leophonte PRole of nocturnal oximetry in screening for sleep apnea syndrome in pulmonary medicine. Study of 329 patients. Rev Mal Respir141997201207 |
| 21. | Rechtschaffen A, Kales A. A manual of standardized terminology techniques and scoring system for sleep stages of human subjects. Washington, DC: US Goverment Printing Office; 1968. Publication No. 204. |
| 22. | Sleep-Related Breathing Disorders in Adults: Recommendations for Syndrome Definition and Measurement Techniques in Clinical Research. The Report of an American Academy of Sleep Medicine Task Force. Sleep 1999;22:667–689. |
| 23. | Department of Radiology. Biological Sciences Division. University of Chicago Hospitals and Clinics. ROCKIT Software. Chicago, IL: Available ftp at random.bsd.uchicago.edu; 1998. |
| 24. | Hanley J, McNeil BThe meaning and use of area under a receiver operating characteristic (ROC) curve. Radiology14319822936 |
| 25. | Harrell F, Lee K, Mark DTutorial in biostatistics: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med151996361387 |
| 26. | Beck J, Shultz EThe use of relative operating characteristic (ROC) curves in test performance evaluation. Arch Pathol Lab Med11019861320 |
| 27. | Efron B, Tibshirani RBootstrap methods for standard errors, confidence intervals, and other measures of statistical accuracy. Stat Sci119865477 |
| 28. | Efron B, Tibshirani R. An introduction to the bootstrap. Monographs on statistics and applied probability 57. New York: Chapman & Hall; 1993. |
| 29. | Yamashiro Y, Kryger MCPAP titration for sleep apnea using a split-night protocol. Chest10719956266 |
| 30. | Quan S, Howard B, Iber C, Kiley J, Nieto F, O'Connor G, Rapoport D, Redline S, Robbins J, Samet J, Wahl PThe Sleep Heart Health Study: design, rationale, and methods. Sleep20199710771085 |
| 31. | Nieto F, Young T, Lind B, Shahar E, Samet J, Redline S, D'Agostino R, Newman A, Lebowitz M, Pickering TAssociation of sleep-disordered breathing, sleep apnea, and hypertension in a large community-based study. Sleep Heart Health Study. JAMA283200018291836 |
| 32. | Young T, Blustein J, Finn L, Palta MSleep-disordered breathing and motor vehicle accidents in a population-based sample of employed adults. Sleep201997608613 |
| 33. | Engleman H, Kingshott R, Wraith P, Mackay T, Deary I, Douglas NRandomized placebo-controlled crossover trial of continuous positive airway pressure for mild sleep apnea/hypopnea syndrome. Am J Respir Crit Care Med1591999461467 |
| 34. | Flemons W, Remmers JThe diagnosis of sleep apnea: questionnaires and home studies. Sleep191996S243S247 |
| 35. | Kump K, Whalen C, Tishler P, Browner I, Ferrette V, Strohl K, Rosenberg C, Redline SAssessment of the validity and utility of a sleep-symptom questionnaire. Am J Respir Crit Care Med1501994735741 |
| 36. | Kupaniai L, Andrew D, Crowell D, Pearce JIdentifying sleep apnea from self-reports. Sleep111988430436 |
| 37. | Haraldsson P, Carenfelt C, Knutsson E, Persson H, Rinder JPreliminary report: validity of symptom analysis and daytime polysomnography in diagnosis of sleep apnea. S151992261263 |
| 38. | Viner S, Szalai J, Hoffstein VAre history and physical examination a good screening test for sleep apnea? Ann Intern Med1151991356359 |
| 39. | Hoffstein V, Szalai JPredictive value of clinical features in diagnosing obstructive sleep apnea. Sleep161993118122 |
| 40. | Kushida C, Efron B, Guilleminault CA predictive morphometric model for the obstructive sleep apnea syndrome. Ann Intern Med1271997581587 |
| 41. | Bradley P, Mortimore I, Douglas NComparison of polysomnography with ResCare Autoset in the diagnosis of the sleep apnoea/hypopnoea syndrome. Thorax50199512011203 |
| 42. | Fleiss J. The design and analysis of clinical experiments. New York: John Wiley & Sons; 1986. |
This article has an online data supplement, which is accessible from this issue's table of contents online at www.atsjournals.org
