Classification algorithms for predicting sleepiness and sleep apnea severity

Abstract

SUMMARY Identifying predictors of subjective sleepiness and severity of sleep apnea are important yet challenging goals in sleep medicine. Classification algorithms may provide insights, especially when large data sets are available. We analyzed polysomnography and clinical features available from the Sleep Heart Health Study. The Epworth Sleepiness Scale and the apnea–hypopnea index were the targets of three classifiers: k-nearest neighbor, naive Bayes and support vector machine algorithms. Classification was based on up to 26 features including demographics, polysomnogram, and electrocardiogram (spectrogram). Naive Bayes was best for predicting abnormal Epworth class (0–10 versus 11–24), although prediction was weak: polysomnogram features had 16.7% sensitivity and 88.8% specificity; spectrogram features had 5.3% sensitivity and 96.5% specificity. The support vector machine performed similarly to naive Bayes for predicting sleep apnea class (0–5 versus >5): 59.0% sensitivity and 74.5% specificity using clinical features and 43.4% sensitivity and 83.5% specificity using spectrographic features compared with the naive Bayes classifier, which had 57.5% sensitivity and 73.7% specificity (clinical), and 39.0% sensitivity and 82.7% specificity (spectrogram). Mutual information analysis confirmed the minimal dependency of the Epworth score on any feature, while the apnea–hypopnea index showed modest dependency on body mass index, arousal index, oxygenation and spectrogram features. Apnea classification was modestly accurate, using either clinical or spectrogram features, and showed lower sensitivity and higher specificity than common sleep apnea screening tools. Thus, clinical prediction of sleep apnea may be feasible with easily obtained demographic and electrocardiographic analysis, but the utility of the Epworth is questioned by its minimal relation to clinical, electrocardiographic, or polysomnographic features.

The Impact of Body Posture and Sleep Stages on Sleep Apnea Severity in Adults

“…We aimed to approximate the distribution of AHI values in the large population Sleep Heart Health Study, with a peak in the normal range followed by a long-tailed distribution extending into the severe range. 12 We assumed that the categories of no apnea (AHI < 5/h), mild apnea (AHI 5-15/h), and moderate apnea (AHI 15-30/h) were normally distributed with mean (SD) values of 2.5/h (1.25), 10/h (2.5), and 22.5/h (7.5), respectively. The severe cohort was modeled using a skew-normal (long-tailed) distribution with a mean AHI of 50/h and 95% of the values between 30/h and 90/h.…”

Covariate Adjusted Receiver Operating Characteristic Curve Analysis for the Evaluation of the Epworth Sleepiness Scale in Diagnosis of Obstructive Sleep Apnea

“…Dolayısıyla zaman ve maliyetten tasarruf etmek ve bekleme sürelerini kısaltmak için polisomnografi öncesi OSA açısından yüksek riskli hastaları belirleyebilmek oldukça önemlidir. Bu konuda klinisyenlere yardımcı olması amacıyla geliştirilmiş birçok tahmin modeli ve ölçek bulunmaktadır (10,11). Bu ölçeklerden biri olan ESS, uyku bozukluklarının değerlendirmesinde sıklıkla kullanılmakta ve OSA için bir tanı testi olarak kullanılabilirliği ile ilgili birçok çalışma bulunmaktadır (3,12).…”

Nocturnal Hot Flashes: Relationship to Objective Awakenings and Sleep Stage Transitions

“…Examples of dissociation between subjective experience and objective measurements are not uncommon in sleep medicine. For example, patients with insomnia may exhibit misperception, 25 patients with sleep apnea may not report daytime sleepiness, 26 and subjects undergoing sleep deprivation may show performance impairment without corresponding subjective insight.…”