Efficient decision making is essential in any venue of healthcare. Many decisions must be made from an abundance of information with limited time and finite resources. The challenge is to decide what information is important, accessible, and accurate for the task at hand. To facilitate decision making, clinical nurse specialists (CNSs) and other healthcare providers use screening questionnaires and other instruments to help focus on important characteristics of a disease or condition and then determine the risks and benefits of different treatment plans. The preoperative assessment is an example of a critical need for efficient decision making. Patients who need surgery may have comorbid conditions, and clinicians may have a limited amount of time to assess, diagnose, and manage their condition. Obstructive sleep apnea (OSA) is a comorbid condition that requires planning, may take considerable time and effort to diagnose, and has significant risks to health. For the CNS working in a surgical clinic, surgical unit, or any environment involving sedation, being aware of a patient's risk of OSA may prevent a catastrophic outcome. The purpose of this article is to present the psychometric properties of the STOP-Bang sleep apnea questionnaire, an instrument designed to help identify OSA in adults prior to surgery.
BACKGROUND
Obstructive sleep apnea is 1 of the most common medical conditions identified in the last 50 years.1 The prevalence differs depending on the population being studied and the diagnostic criteria used to define OSA. But overall, OSA ranges from 2% to 26% in the general population, with more than 80% of patients being undiagnosed.2
Obstructive sleep apnea is characterized by intermittent and recurrent episodes of partial or complete upper airway obstruction during sleep, resulting in fragmented sleep and daytime sleepiness.3 Obstructive sleep apnea is associated with obesity, coronary artery disease, hypertension, and diabetes.4 Specific to surgical patients with OSA, medications for sedation, anesthesia, and pain control may increase upper airway collapsibility by decreasing the muscle activity of the genioglossus and worsening OSA.3 The combination of comorbid diseases and sensitivity to medications used for anesthesia, sedation, and pain management creates a particular hazard of worsening the obstruction and resulting apnea. If the diagnosis of OSA is known or suspected, a different plan of care should be implemented. For example, utilizing regional anesthesia, minimizing narcotics and sedatives, and preparing for additional monitoring and support in the postoperative period would be appropriate. Same-day surgery patients may need to be admitted overnight, procedural sedation in the outpatient clinic may not be appropriate, continuous positive airway pressure (CPAP) machines may need to be brought in with the patient, and judicious use of patient controlled analgesia and continuous respiratory monitoring may be required.
As always, every interaction with a patient is an opportunity to assess his/her health knowledge, provide education, or reinforce important health teachings-core functions of nurses that are supported by the CNS role. The key to this process concerning OSA patients is the knowledge or clinical suspicion of OSA prior to the administration of any anesthetic, narcotic, or sedative. The diagnosis of OSA is the challenge.
Historically, the criterion standard for establishing the diagnosis of OSA is to obtain an overnight polysomnography sleep study. As previously stated, there is a high incidence of OSA in the general population, and attempting to screen all patients with polysomnography would be burdensome and cost-prohibitive. In addition, there are problems with patient acceptance of treatment options such as losing weight or use of airway modalities, such as CPAP, and patients may defer participating in a sleep study and go undiagnosed.
Over time, many different instruments and processes have been purported to screen for OSA ranging from simple questionnaires to complex computer programs. A recent meta-analysis found 8 questionnaires, 18 regression models, neural networks, and algorithms that could be classified as clinical prediction tests for OSA.5 However, the majority of these instruments and processes were designed for use in a sleep clinic and are not well suited for the busy environment of the preoperative area in a hospital or clinic. Based on the unique needs of the anesthesia preoperative environments and surgical patients, an efficient screening instrument was needed. A screening instrument that is reliable, produces valid data for diagnosis, and is easy to use could significantly alter the plan of care for patients with OSA as well as decrease risks for negative outcomes.
DEVELOPMENT OF STOP-BANG SLEEP APNEA QUESTIONNAIRE
The STOP-Bang questionnaire was developed by Chung et al6 based on clinician-raised issues encountered with the Berlin questionnaire, described in the following paragraphs. Revisions to the Berlin questionnaire led to the STOP questionnaire, and that questionnaire was later refined into the STOP-Bang questionnaire.
Berlin Questionnaire. At the time of the development of the STOP-Bang, the Berlin questionnaire had been the most widely used screening instrument for OSA. The 11 Berlin questionnaire items were categorized into (a) snoring, (b) sleepiness/fatigue, and (c) high blood pressure categories.6 However, the scoring system was cumbersome in that each item could have multiple positive responses such as "score item 2 as 'c' or 'd' and assign 1 point." Next, the items and their respective categories had to be assessed as "positive" or "negative," and finally, the number of positive categories is determined. If 2 or more categories are scored positive, a patient is determined to be at risk of OSA.7 This process was not at all efficient for busy clinicians.
In both primary care and postoperative care patients, the Berlin questionnaire sensitivities (similar to reliability testing and having the ability to determine someone has OSA) ranged from 54% to 86%, and specificities (the ability to rule someone out who does not have OSA) ranged from 77% to 97%.6,8 The lower range for sensitivity was in the primary care population, indicating that the instrument was more sensitive when used with postoperative patients.
STOP Questionnaire. Chung et al9 initially moved to develop the STOP questionnaire (Bang was added later) based on the Berlin questionnaire items, a review of scientific literature, and input from a group of anesthesiologists and sleep specialists. As a result, the STOP questionnaire uses a simple mnemonic and can be self-administered.9
In addition to 7 demographic items, 4 items for the STOP questionnaire were eventually selected from a total of 14 items by the professional group. Factor analysis on 14 initial items resulted in the following factors: (1) S-snoring, (2) T-tiredness, (3) O-observed as stopped breathing during sleep, and (4) P-treatment for high blood pressure. To simplify the instrument, the healthcare professionals chose 1 item to represent each factor. Responses for each of the final 4 items are recorded as yes or no. Yes answers to 2 or more questions indicate that the patient would be at high risk of OSA. In addition, the instrument included 7 demographic items including height, weight, age, gender, body mass index, collar shirt size (or shirt size measurement), and neck circumference measured by staff.9
Reliability of the STOP questionnaire has been determined using test-retest agreement. In this case, the STOP questionnaire was administered twice to 55 patients, 1 to 27 days (median, 8 days) apart. The majority (96%) of those patients had the same score (K coefficient of 0.923; confidence interval, 0.82-100), indicating that the STOP questionnaire was stable over time. Predictive validity for the STOP questionnaire has been determined using a sample of patients (n = 211) who volunteered for a sleep study. Those determined to have OSA using the STOP questionnaire had apnea-hypopnea indices (AHIs) greater than 5 (P < .05).9 The AHI is a useful comparison for predictive validity. It is a summation of the number of apneas and hypopneas (decreased breathing with oxygen desaturation) per hour of sleep, having been used for quantifying disease severity in numerous settings.10
DESCRIPTION OF THE STOP-BANG QUESTIONNAIRE
The prevalence of OSA is higher in the surgical population than the general population and varies within different subgroups of the surgical population, such as preoperative patients. The STOP questionnaire had been utilized in primary care and with postoperative patients but less so with preoperative patients. Therefore, the STOP questionnaire underwent additional testing to determine its ability to generate valid data in preoperative settings. In addition, there was a desire to improve the sensitivity of the STOP questionnaire so that it could better identify patients with moderate OSA. Therefore, Chung and colleagues9 made a structural change to the STOP questionnaire and renamed the new version STOP-Bang.
The STOP-Bang questionnaire has a total of 8 items all scored by circling either yes or no. The first 4 items are from the STOP questionnaire (S-"Do you snore loudly [louder than talking or loud enough to be heard through closed doors]?" T-"Do you often feel tired, fatigued, or sleepy during the daytime?" O-"Has anyone observed you stop breathing during your sleep?" P-"Do you have or are you being treated for high blood pressure?"). The final 4 items (the new Bang portion) are based on 4 demographic items chosen from the STOP questionnaire (b-body mass index >35 kg/m2, a-age >50 years, n-neck circumference >40 cm, and g-gender [male]). Similar to the first 4 items, responses are recorded as either yes or no.9 The newly designed instrument can be completed in less than 5 minutes and provides a simple end point of risk assessment for OSA.
Scoring
All 8 items have forced choice yes/no responses, and scores range from 0 to 8. A scale and a tape measure are required to determine body weight and neck circumference. The instrument units of measure for body weight and neck circumference are metric but Imperial or US Standard units can easily be converted to metric. Yes responses to 3 or more items indicate a high risk of OSA; yes responses to less than 3 items are a low risk of OSA. While being able to identify a high- and a low-risk category initially, instrument reliability testing (discussed next) assisted in identifying a middle range for OSA.
Reliability
The desire to have an instrument that could be reliable when measuring smaller increments (not just AHIs >5) resulted in testing the STOP-Bang for a range of sensitivities. The STOP-Bang sensitivities ranged from 83.6% for AHI of greater than 5, 92.9% for AHI of greater than 15, and 100% for AHI of greater than 30. These test results indicate that the STOP-Bang is sensitive to identifying OSA over a range of AHIs, including patients with low, moderate, and severe OSA. Chung et al9 directly attributed this gain in sensitivity to the addition of the body mass index, age older than 50 years, neck circumference of more than 40 cm, and gender (male) into the scoring scheme of yes/no responses.
Validity
According to Chung et al,6 higher scores on the STOP-Bang were associated with the occurrence of postoperative complications. This indicates that the STOP-Bang has some predictive properties. Validity testing in 2012 demonstrated that higher scores in the range of 7 to 8 on the STOP-Bang were indicative of moderate to severe OSA.11 Difficult intubations have been correlated with STOP-Bang scores of 3 or greater.12 Among obese patients, STOP-Bang scores identified patients at risk for difficult airways to manage (P < .001).13
A number of studies have been conducted to determine validity of data produced by the STOP-Bang in non-English-speaking patient populations. For example, Arabic and Persian versions have been determined to classify patients at risk of OSA.14-16
RECOMMENDATIONS
In clinical practice, a positive screen for OSA signals that routine doses of narcotics, sedatives, and anesthetics may have a stronger effect. A standard dose of a medication before surgery may lead to significant sedation. A usual infusion rate for sedation may leave the patient obtunded and apneic. Long-acting medications may significantly delay emergence, extubation, and recovery. Respiratory support in the form of supplemental oxygen, airway adjuncts, CPAP, and end-tidal CO2 monitoring may be required. For outpatients, discharge to home may be significantly delayed. Finally, for inpatients, an unplanned admission to a higher level of care may occur with patients who have OSA.
Designed with an easy mnemonic, the STOP-Bang questionnaire was developed for screening surgical patients, including preoperative patients, for the presence of OSA. Previous screening tools were cumbersome to use, complicated to score, and time consuming and had not been validated in the more specific surgical populations. STOP-Bang has been demonstrated to be efficient and reliable, producing valid assessments in the surgical patient population to screen for undiagnosed OSA. High STOP-Bang scores have been associated with postoperative complications9 and therefore can be used by CNSs to identify high-risk patients.
Across all 3 instruments presented here, a significant amount of psychometric testing has been completed. However, additional psychometric testing of the STOP-Bang questionnaire is recommended. While previous versions of the STOP-Bang have undergone testing, these results should not be assumed to substantiate the reliability of this latest version. In addition, there are a number of patient populations for which there had been no testing to determine validity. For example, pregnant women are a unique subgroup of the surgical population who has been understudied for OSA. The core factors of STOP-Bang were not developed for the dynamic process of pregnancy. Tantrakul et al17 recently evaluated the STOP-Bang utilizing at-home, overnight, sleep testing throughout pregnancy. The STOP-Bang was found to be of limited usefulness in predicting OSA across the trimesters and cited the dynamic process of pregnancy, specifically the second trimester, as uniquely challenging.
Information on how the STOP-Bang performs in diverse patient populations is also limited. While Arabic and Persian versions have undergone validity testing, other language versions remain untested.
In summary, the STOP-Bang questionnaire uses a simple mnemonic (STOP-Bang) to identify the instrument items, is simple to score, and has some demonstrated reliability and predictive validity. Further psychometric testing is recommended to establish stability using test-retest procedures and consistency of ratings such as interrater and intrarater reliability testing. Additional testing is also recommended across different ethnic and language-speaking patient populations to determine the validity of scores produced by the STOP-Bang so that CNSs and other healthcare professionals can be confident in using the instrument in diverse patient populations.
References