A New Berlin Questionnaire Simplified by Machine Learning Techniques in a Population of Italian Healthcare Workers to Highlight the Suspicion of Obstructive Sleep Apnea

Abstract

Obstructive sleep apnea (OSA) syndrome is a condition characterized by the presence of repeated complete or partial collapse of the upper airways during sleep associated with episodes of intermittent hypoxia, leading to fragmentation of sleep, sympathetic nervous system activation, and oxidative stress. To date, one of the major aims of research is to find out a simplified non-invasive screening system for this still underdiagnosed disease. The Berlin questionnaire (BQ) is the most widely used questionnaire for OSA and is a beneficial screening tool devised to select subjects with a high likelihood of having OSA. We administered the original ten-question Berlin questionnaire, enriched with a set of questions purposely prepared by our team and completing the socio-demographic, clinical, and anamnestic picture, to a sample of Italian professional nurses in order to investigate the possible impact of OSA disease on healthcare systems. According to the Berlin questionnaire, respondents were categorized as high-risk and low-risk of having OSA. For both risk groups, baseline characteristics, work information, clinical factors, and symptoms were assessed. Anthropometric data, work information, health status, and symptoms were significantly different between OSA high-risk and low-risk groups. Through supervised feature selection and Machine Learning, we also reduced the original BQ to a very limited set of items which seem capable of reproducing the outcome of the full BQ: this reduced group of questions may be useful to determine the risk of sleep apnea in screening cases where questionnaire compilation time must be kept as short as possible.

Obstructive Sleep Apnea: Respiratory Complications – A Review Article

“…Table (1) shows relevant studies on the association between OSA and perioperative outcomes. A number of factors, including age, American Society of Anesthesiologists (ASA) class, and medical history, raise a patient's risk of problems after surgery, as well as other factors; type of paralytics, low albumin, current smoking, duration of surgery, type of anaesthesia, and other comorbidities, particularly coronary artery disease (CAD), chronic obstructive pulmonary disease (COPD), and renal failure (27) .…”

“…Disruptions to the sleep cycle, such as these, lead to tiredness during the day. Obstructive sleep apnea syndrome (OSAS) has been linked to an increased risk of car accidents, diabetes, hypertension, heart disease, stroke, and overall mortality (1) .…”

Machine Learning (ML) Models to Enhance the Berlin Questionnaire (BQ) Detection of Obstructive Sleep Apnea (OSA) at-Risk Patients

“…Recent advancements in software technologies and Machine Learning (ML) methods have significantly enhanced the development of effective predictive and diagnostic tools, becoming increasingly prevalent in various fields of medical research and applications, including for OSA [12,13,[21][22][23][24][25][26][27][28]. The prediction models described in existing research primarily utilize clinical data, such as demographic information (age and gender), comorbid conditions, anthropometric measures (Body Mass Index (BMI), waist and neck circumferences), symptoms of OSA, and physiological parameters (blood pressure, overnight pulse oximetry, and lung function tests).…”

“…Originally introduced in the United States (US), the BQ consists of a concise set of questions focused on risk factors and symptoms associated with OSA, aimed at identifying patients at high risk who might benefit from undergoing PSG to facilitate increased diagnosis rates. While the standard BQ comprises 10 questions, we previously introduced a streamlined version by using a trained classifier, known as the simplified Berlin questionnaire (BQ_2) [23], which reduces the questionnaire to just two questions. This abbreviated version has been shown to achieve results comparable to the original BQ, offering an efficient means of screening highrisk OSA patients.…”

“…For this purpose, we developed a ML-enhanced BQ model (ML-10) capable of predicting the risk of OSA using the same items as the BQ. Additionally, we explored a simplified version of ML-10, called BQ-2 [23], which is based on the BQ but only preserves two of the original BQ items, to determine whether it yields comparable results. The predictive performance of these models was evaluated against the conventional BQ approach, which does not incorporate ML techniques.…”