A protocol for mitigating safety events in a sleep laboratory

Abstract

Study Objectives:

Polysomnography is a common outpatient procedure and the rate of adverse events is considered low. Due to the emergence and use of home sleep apnea testing, the patient population presenting for in-laboratory testing may have greater medical complexity, suggesting greater risk for in-laboratory adverse events. We believe that there is a greater need for standardized protocols to triage medically vulnerable populations and for formalized training of sleep technicians to respond to safety events.

Methods:

The sleep laboratories affiliated with the Beth Israel Deaconess Medical Center system developed a referral triage protocol for patients undergoing polysomnography and a training protocol for sleep technicians with a formalized response to medical incidents. Safety events occurring from January 2016 to January 2020 were documented and patient demographics, referral characteristics, event characteristics, and outcomes were analyzed.

Results:

Sixty-five safety events occurred over this period, with a rate of 1:147 studies. The most common events were chest pain (20/65, 31%), shortness of breath (13/65, 20%), and vital sign abnormalities (12/65, 18%). Patients experiencing events were 49% (32/65) female, with a median age of 57 years (range, 19–91 years); 60 of 65 (92%) had documented medical comorbidities, with a median of 3 documented medical or psychiatric comorbidities (range, 0–9). With the formalized response protocol, the time from incident identification to activation of the appropriate response was a median of 3 minutes (range, 0–47 minutes).

Conclusions:

The incidence of in-laboratory safety events may be greater than previously described due to the widespread use of home sleep apnea testing. Implementation of formalized response protocols and sleep technician training may be necessary to meet the needs of an increasingly medically complex population.

Citation:

Blattner M, Dunham K, Thomas R, Ahn A. A protocol for mitigating safety events in a sleep laboratory. J Clin Sleep Med. 2021;17(7):1355–1361.

BRIEF SUMMARY

Current Knowledge/Study Rationale: Due to the emergence and use of home sleep apnea testing, the patient population presenting for in-laboratory testing may have greater medical complexity, suggesting greater risk for in-laboratory adverse events. We believe there is a greater need for standardized protocols to triage medically vulnerable populations and to train sleep technicians to respond to safety events. This study describes a generalizable protocol for triage and training and describes the breadth of safety incidents, responses, and outcomes in a sleep laboratory.

Study Impact: Safety events occurred at a rate of 1:147. The most common events were chest pain, shortness of breath, and vital sign abnormalities. These findings suggest that the incidence of in-laboratory safety events may be greater than described before widespread use of home sleep apnea testing. Implementation of safety event response protocols and sleep technician training may be necessary to meet the needs of a medically complex population.

INTRODUCTION

Polysomnography is a common overnight outpatient procedure requiring continuous cardiopulmonary monitoring by a trained technologist in a sleep laboratory. The traditional clinical sleep laboratory has been a place for noninvasive assessments of sleep disorders in relatively healthy individuals. Although safety risks are present,¹ prior studies have indicated an exceedingly low adverse event rate of 0.35% or 0.16%, where most adverse events are either cardiac in nature or related to falls.^2,3 These prior observational studies predate the advent of home sleep apnea testing and prior authorization requirements and may not reflect the current population now presenting for hospital-based sleep evaluation.

With the advent of prior authorization and home sleep apnea testing, the majority of relatively healthy adults seeking a sleep study evaluation are referred for home sleep apnea testing. It has been our observation that patients approved for in-laboratory sleep studies are older, require greater mobility assistance, and have significant neurological or cardiovascular comorbidities. As patients in sleep laboratories have increased medical and psychiatric comorbidities and require greater support, the risk for in-laboratory incidents may also increase.

Presently, sleep technicians are trained to apply sensors, monitor sleep patterns, and apply positive airway pressure therapy when warranted. Their training does not necessarily prepare them to quickly evaluate a patient with acute symptoms and appropriately determine the need for urgent medical evaluation. Given the increased clinical complexity of patients referred for laboratory-based studies and the minimal emergency training of sleep technicians, we identified a need for an explicit triage protocol for our laboratory-based sleep studies including more formalized cardiopulmonary training for our sleep technicians. Additionally, isolated sleep laboratory safety events both at our sleep laboratory and nationally have demonstrated a critical need for formalized assessment and response protocols. This paper describes our approach to implementation of a comprehensive sleep laboratory safety protocol through triage of referrals, sleep technician training, and a formalized incident response protocol. We also describe the rate, characteristics, and outcomes of safety incidents observed in our sleep laboratory.

METHODS

In-laboratory sleep study location and resources

The Multi-Disciplinary Sleep Disorders Center at the Beth Israel Deaconess Medical Center (Boston, MA) utilizes 3 Massachusetts locations for in-laboratory sleep studies: one 8-bed hospital-based sleep laboratory (Beth Israel Deaconess Hospital Needham), one 2-bed hospital-based sleep laboratory (Beth Israel Deaconess Hospital Milton), and one 4-bed laboratory at a non-hospital-based affiliated sleep center. The Beth Israel Deaconess Hospital Needham and Beth Israel Deaconess Hospital Milton hospital-based sleep laboratories are located adjacent to the emergency department (ED) and have 24-hour in-house hospitalist coverage with rapid response (RR) and code teams. Both hospital-based sleep laboratories and the non-hospital-based laboratory are managed by an affiliated sleep center, NeuroCare Center for Sleep, in Newton, Massachusetts.

Prior workflow and needs assessment

Our needs assessment was motivated by isolated sleep laboratory safety events both at our sleep laboratory and nationally. The need for a safety process was highlighted by the media announcement in the fall of 2015 of the wrongful death case at Emory Sleep Center in 2010 of a 25-year-old man who developed hypoxia and respiratory arrest during his sleep study.⁴ An additional incident in our own sleep laboratory also in the fall of 2015 involved a patient with history of atrial fibrillation who developed chest pain and tachycardia and required transfer to an ED for management of atrial fibrillation with rapid ventricular rate. At that time, the protocol involved calling an intermediary nursing supervisor, and delays in that response resulted in the sleep laboratory technician transferring the patient to the ED directly.

Together, these situations motivated the development of clearer triage guidelines and training for our technicians in patients with acute reported symptoms (such as chest pain, shortness of breath) or abnormal critical findings (such as low oxygen saturation, low or high blood pressure, fast or irregular heart rate).

Additional protocols were created in response to potentially high-risk situations: for example, the development of explicit prescreening, arrival, and monitoring procedures for any patient with a tracheostomy. The tracheostomy protocol was prompted by an incident where a patient with congenital central hypoventilation syndrome and tracheostomy became transiently hypoxemic when placed on different ventilator settings. He recovered quickly when his home ventilator settings were re-established. Another formal protocol was put in place for use of sedating medications. This arose from an isolated event in 2016 involving a patient taking zolpidem for the first time at study initiation, and then unexpectedly leaving in the middle of the study. Although the patient was felt to be alert and capable of making decisions, the next morning the patient’s partner informed the laboratory that the patient had been found asleep in the car in the driveway without a recollection of leaving the laboratory or driving home. Following a critical review of this event, a detailed sedative use policy was created.

Prior to 2015 and the implementation of our protocol, management of acute symptoms during a laboratory-based sleep study included instructing the laboratory technician to contact the nursing supervisor and the on-call sleep medicine physician for patient complaints (shortness of breath, headache, or chest pain). There was no formal prestudy risk-stratification questionnaire or comprehensive prestudy vital sign measurement. No specific sedation or tracheostomy policy was in place. The RR and code teams were not explicitly involved by protocol.

To assess the skill level and comfort with responding to patient emergencies of the sleep technicians, an emergency workshop was held. The workshop incorporated patient simulations for emergency situations, such as chest pain, arrhythmias, and respiratory distress. The results from that workshop were unanticipated and surprising: Most technicians were not comfortable identifying what symptoms or complaints should prompt a call for emergent medical support.

Multistep, multicomponent, parallel-process modification

A number of organizational changes were made to the emergency protocols including removing the nursing supervisor from the emergency response protocol and instead utilizing the RR and code teams, creating an annual emergency training program for the sleep technicians, obtaining a code cart for the sleep laboratory hallway and manual blood pressure cuffs for sleep laboratory use, posting “Take Quick Action” forms in sleep laboratory and control rooms, and reviewing every incident for response appropriateness and choosing 1–2 incidents each year to complete a root-cause analysis.

Sleep laboratory referral process and study location triage (part 1)

The vast majority (> 95%) of polysomnograms are performed in the hospital-based sleep laboratory. The electronic sleep study order was modified to explicitly require referring physicians to include any special needs, such as mobility limitations, cognitive impairment, insulin requiring diabetes, seizure history, or use of supplemental oxygen (Box 1). The presence of 1 of these factors or conditions on the electronic order prompted triage to the hospital-based Beth Israel Deaconess Hospital Needham sleep laboratory, which was identified as having increased medical support. The electronic sleep study order was also modified to direct any patient identified as needing adaptive servoventilation to the sleep clinic prior to scheduling the sleep study. External referrals were reviewed for the same criteria before scheduling. This modification to the referral process preceded the data collection reported here.

Box 1.

High-needs criteria listed in the electronic sleep study referral for triaging to a hospital-based sleep laboratory (presence of ≥ 1 criteria).

Mobility limitations or additional assistance

Wheelchair or walker

Group living facility

Incontinent

Need an attendant to accompany

Visual impairment

Medical history

Recent hospitalization

Home nocturnal oxygen requirement

Insulin required for diabetes

Cognitive impairment

Neuromuscular disease

History of seizures

Morbid obesity

Tracheostomy present

Class III or IV congestive heart failure

Cardiac arrhythmia

History of myocardial infarction or stroke in past 3 months

Moderate to severe pulmonary disease

Sedative use policy

A sedative use policy was created and included in the prestudy paperwork sent to patients. This policy explicitly instructed patients to take their sedative medication by 10:30 pm and stated that if a sedative was to be used for the first time during the sleep study, then the patient was encouraged to arrange a ride home the morning following the study.

Engaging the hospitalists and the RR and code teams

Implementation of protocols was coordinated with the hospital’s RR and code teams to delineate responsibilities of the RR or code teams and the sleep technicians. This included designating a location and maintaining a code cart within the sleep laboratory.

Clinical symptom screening and triage upon arrival to the sleep laboratory (part 2)

A clinical symptom screening and triage protocol was developed, tested, and refined between 2016 and 2017. Upon patient arrival to the sleep laboratory, the technician performed prestudy vital signs assessment using an automated blood pressure cuff and oximeter and administered a prestudy clinical symptom questionnaire. Acceptable baseline blood pressure and oximetry ranges as well as actions for vital signs that are out of range were explicitly defined in a “Take Quick Action” triage sign (Figure S1 ^{(2.4MB, pdf)} in the supplemental material). The prestudy questionnaire identified high-risk conditions as well as acute symptoms within the past 24 hours (Figure S2 ^{(2.4MB, pdf)} in the supplemental material). If the patient answered “yes” to certain questions on the prestudy questionnaire, then the technician was prompted to call the on-call sleep medicine physician, the RR team, or the code team. The technician also explicitly discussed the sedative medication policy with the patient.

Emergency protocol revision and training of sleep technicians (part 3)

A “Take Quick Action” triage sign was developed that explicitly described the appropriate response for each incident (Figure S1 ^{(2.4MB, pdf)} ). Sleep technicians were instructed to call the RR team for chest pain, shortness of breath, fall, seizure, stroke symptoms, voiced intent for self-harm, marked concern by patient or family member, or vital sign abnormalities (heart rate > 120 beats per minute for 2 minutes, wake SpO₂ < 90%, systolic blood pressure < 90 mm Hg or > 200 mm Hg). Sleep technicians were instructed to call the code team for cardiac arrest or difficulty arousing the patient. Other concerns were directed to the sleep physician on call. All incidents were reported to the sleep physician in the morning and communicated to the referring clinician.

Emergency training was implemented for all sleep technicians and included an annual review of emergency procedures and an annual review of electrocardiographic tracings and causes of respiratory distress, with the creation of an advanced electrocardiographic skills training program; the creation of a script for how technicians should respond in an emergency with a review of scenarios; the creation of a specific tracheostomy prestudy form (Figure S3 ^{(2.4MB, pdf)} in the supplemental material), and the creation of a sedative use protocol (Figure S4 ^{(2.4MB, pdf)} in the supplemental material).

Incident definition and tracking outcomes

Incidents were defined as situations in which the sleep technician was required to evaluate an unexpected situation or symptom and initiate a treatment plan to mitigate or minimize risk or harm. These included events with a phone call to emergency personnel (the RR team or code team), a phone call for security assistance for behavioral dysregulation, a phone call to the sleep physician on call for assistance concerning dangerous decision-making by the patient (leaving the study prematurely after taking a sedative-hypnotic, after an incident occurred, or after a new serious symptom or arrhythmia had been identified), or any injury to the patient during the time in the sleep laboratory (not including minor skin bruising or irritation during routine placement of electrodes). Incident characteristics were tracked and reviewed every 3–6 months to evaluate the need for updated protocols.

RESULTS

Patient demographics, medical comorbidities, and referral characteristics

Patient demographic details, medical and psychiatric comorbidities, and study referral information are presented in Table 1. Incidents occurred equally in males and females (49% [32/65] female) with a median age of 57 years (range, 19–91 years). Over 90% of the patients with reported incidents had some medical or psychiatric comorbidity, with a median of 3 documented comorbidities (range, 0–9). Of the incidents documented, 64 of 65 (98%) occurred in the hospital-based sleep laboratories, rather than in the non-hospital-based sleep laboratory. The 1 incident reported in the non-hospital-based sleep laboratory was a complaint of chest pain in a 30-year-old woman with a history of myotonic dystrophy and depression. Recorded vital signs were stable during the event; 911 was called, and the symptoms resolved after talking with the emergency personnel.

Table 1.

Characteristics of patients and sleep referrals in reported safety events.

	Finding
Patient characteristics
Median (range) age, y	57 (19–91)
Female sex, n (%)	32 (49)
Race, n (%)
White, non-Hispanic	42 (65)
Black or African American	11 (17)
Hispanic	5 (8)
Other	1 (2)
Not documented	6 (9)
Comorbidities,a n (%)
Any	60 (92)
Cardiac	35 (54)
CHF	8 (12)
Atrial fibrillation	8 (12)
Neurologic	24 (37)
Stroke	1 (2)
Parkinson disease	4 (6)
Psychiatric	24 (37)
Endocrine	20 (31)
Pulmonary	16 (24)
COPD	7 (11)
Asthma	4 (6)
Rheumatologic	3 (5)
Number of comorbidities, median (range)	3 (0–9)
Referral characteristics, n (%)
Referral
From sleep clinician	44 (68)
From non-sleep clinician	21 (32)
Study type
Split polysomnogram	31 (48)
Diagnostic polysomnogram	21 (32)
Titration polysomnogram	12 (18)
Not documented	1 (2)
Suspected diagnosis
Obstructive sleep apnea	58 (89)
Insomnia	2 (3)
Central sleep apnea	1 (2)
Parasomnia	1 (2)
REM sleep behavior disorder	1 (2)
None documented	1 (2)

n = 65. ^aDocumented comorbidities: cardiac (hypertension, atrial fibrillation, coronary artery disease, CHF), pulmonary (asthma, restrictive lung disease, chronic obstructive lung disease, interstitial lung disease, cystic fibrosis), neurologic (dementia, history of stroke, epilepsy, pseudotumor, traumatic brain injury, Parkinson disease, occipital neuralgia, myotonic dystrophy, meningioma, congenital central hypoventilation syndrome), rheumatologic (rheumatoid arthritis, autoimmune hepatitis, psoriasis, fibromyalgia), psychiatric (depression, psychosis, bipolar, posttraumatic stress disorder, attention-deficit disorder, attention-deficit/hyperactivity disorder, substance abuse, developmental delay), and endocrine (diabetes, prediabetes, hyperparathyroidism, polycystic ovarian syndrome, hypothyroidism). CHF = congestive heart failure, COPD = chronic obstructive pulmonary disease, REM = rapid eye movement.

Incident reporting

After implementation of the new incident response protocols, the number of reported events increased from 6 reports in 2015 to 22 reports in 2016 (annual reports from 2016–2019 ranged from 11–22/year). Over the study period of January 2016–January 2020, there were 9,558 total in-laboratory polysomnograms completed. With 65 events reported over this time, the incidence of safety incidents was 1:147 for all in-laboratory sleep studies performed.

Incident characteristics and outcomes

The most common events were symptoms of chest pain (20/65, 31%) and shortness of breath (13/65, 20%) (Table 2). Technician-documented vital sign abnormalities were also noted (12/65, 18%). The action taken in response to incidents was based on the formalized protocol and was most often activation of the RR team (40/65, 62%). For the incidents with documented response time, the interval between incident identification and response activation was 3 minutes (range, 0–47 minutes). The median interval between the response activation (ie, calling the RR team) and the intervention (ie, evaluation by a physician) was 3 minutes (range, 0–25 minutes). Most patients were transferred to the ED for evaluation (41/65, 63%), and of these, 51% (21/41) either had sufficient sleep data prior to transfer, many with symptomatic complaints on awakening in the morning (14/21), or completed the study after ED evaluation and incident resolution (7/21). The patients who returned to the study after ED evaluation included 4 patients with acute symptom complaints on arrival triage (chest pain, headache, dizziness) and 1 patient with vital signs outside acceptable parameters on arrival. These patients were evaluated, treated, and discharged from the ED; with all patients, this was completed in < 2 hours and patients were allowed to return to the sleep laboratory for their study. The remaining 2 patients reported dizziness during the study (10:00 pm–midnight), were transferred to the ED, and returned to the sleep laboratory when discharged from the ED (eg, following juice administration for low blood sugar).

Table 2.

Incident characteristics, time to response, and outcomes.

	Finding
Incident characteristics (n = 65)
Patient-reported symptoms, n (%)
Chest pain	20 (31)
Shortness of breath	13 (20)
Headache	8 (12)
Dizziness	5 (8)
Nausea	2 (3)
Technician-noted events, n (%)
Suspected seizure	5 (8)
Syncope	3 (5)
Disorientation	3 (5)
Refractory coughing	1 (2)
Vital sign abnormalities, n (%)
Tachycardia (> 120 beats/min)	5 (8)
Hypoxia (< 88% SpO2) during wake	3 (5)
ECG abnormalities (heart block)	1 (2)
Hyper/hypotension (systolic pressure < 90 or > 200 mm Hg)	3 (5)
Action taken, (> 1 action, in some incidents)
Rapid response team activated	40
Called sleep attending physician on call	11
Called hospitalist on call	8
Code activated	3
Called security	3
Technician brought patient directly to ED	2
No further action	1
Time from incident identification to response activation
Median (range), min	3 (0–47)
Not documented, n	15
Time from response to intervention (arrival of physician or evaluation)
Median (range), min	3 (0–25)
Not documented, n	27
Outcome/disposition, n (%)
Admitted to the ED	41 (63)
Home (against medical advice)	4 (6)
Admitted to hospital floor	2 (3)
Discharged home without completing study	1 (2)
Not documented	2 (3)
Study completed,a n (%)	37 (57)
Study characteristics, completed studies (n = 37)
AHI 3%, median (range), events/h	46.2 (0–134)
AHI 4%, median (range), events/h	19.1 (0–128)
Mean O2, median (range), %	94 (80–97)
Minimum O2, median (range), %	83 (50–92)

^aAdequate testing completed before incident or able to complete the study following the incident. AHI = apnea-hypopnea index, ECG = electrocardiogram, ED = emergency department.

Of the 40 events for which the RR team was paged, most were transferred to the ED (30/40, 75%). Of the patients who were not transferred to the ED, the RR team evaluation provided medical guidance that allowed for completion of the study. In 5 instances, patients had abnormal vital signs without any symptoms (2 with elevated blood pressure, 2 with hypoxia, 1 with tachycardia) and on review of medical history and discussion with the ordering physician, these studies proceeded. Two additional patients with a history of diabetes were found to have hypoglycemia as a cause for their complaints of lightheadedness or confusion, which improved with a snack. The remaining 3 patients complained of chest pain and/or shortness of breath, and again on review of symptoms, vital signs, and clinical history, the RR team recommended completion of the study with improvement of symptoms.

The completed studies for patients with incidents suggested a tendency toward severe sleep apnea, with a median apnea-hypopnea index of 46.2 events/h (range, 0–134 events/h), although there was considerable variability in sleep-disordered breathing in these patients (Table 2).

DISCUSSION

A process for evaluating and managing urgencies and emergencies in a traditional sleep laboratory is described. This approach includes triage of referrals, sleep technician training, and creation of a formalized incident response protocol. While these results suggest that serious medical complications during sleep studies are rare, preparedness and standardized response protocols allow for swift identification and activation of emergency response procedures when they do occur.

Overall events were rare, with 65 events over 3 years (9,558 studies), with a rate of 1:147; however, this incidence was higher relative to previously reported rates³ and may reflect increased patient acuity over time. Patients undergoing in-laboratory polysomnography have shown an increasing number of medical comorbidities over the past 10 years⁵ that likely contributes to the increased incidence of acute events in the sleep laboratory. Colaco et al⁵ further suggest a “Polysomnogram Clinical Index” based on comorbidities as a way to anticipate needed services of more complex patients. Our center does not formally measure the Polysomnogram Clinical Index score, but essentially operationalized it such that anyone with a score of > 1 (or ≥ 1 high-risk criteria [Box 1]) is triaged to our hospital-based sleep laboratory. Because our hospital-based laboratories have 10 beds and our free-standing laboratory has 4 beds, the majority of our sleep study population is able to be served in the hospital-based laboratory, allowing us to reserve the free-standing sleep laboratory beds for those with effectively a Polysomnogram Clinical Index score of 0 or no high-risk criteria.

Previous studies have reported falls during overnight studies. Kolla et al³ report that 5 out of the 12 patients who experienced a fall were given zolpidem by their sleep physician. Those authors shared that at their sleep laboratory, sleep technicians did not receive the same fall-mitigation education that inpatient nurses receive, and that patients and families were not routinely educated in the sleep center about fall risks during sleep study testing. In our center, we did not have any falls reported during the study period, possibly due to the rate of sedative use and the inclusion of fall risk in the triage process. Between January 2016 and December 2020, 548 patients took zolpidem on the night of their study, accounting for approximately 5% of the total number of patients evaluated in the laboratory. Furthermore, any increased risk associated with zolpidem was likely mitigated by prestudy fall-risk screening and the use of attendants in individuals identified as high risk. Any patient identified as having mobility limitations during scheduling or prestudy screening was asked to be accompanied by an attendant during the night of the study.

While sleep technicians typically do not have advanced medical training or Advanced Cardiac Life Support certification, a protocol that includes clear triage and contact instructions can ensure that patients have access to timely evaluation and care. Events generally occurred in patients with multiple medical comorbidities, suggesting that a more complex in-laboratory patient population may require additional screening and support. For example, our center has a program for screening patients with atrial fibrillation. This has resulted in an increase in patients with this arrhythmia and typical associated comorbidities (heart failure, stroke) presenting to the sleep laboratory. If treating sleep apnea in heart failure becomes a clinical standard, then the complexion of the laboratory will change in predictable ways.

These results are generalizable at least to hospital-based systems, while independent testing facilities may see less severely ill patients. Our sleep laboratory has the luxury of in-hospital and free-standing sleep laboratory resources, allowing triaging higher-risk patients to the hospital setting. Independent diagnostic testing facilities that offer sleep testing services may not have this option and may benefit from affiliating with a hospital-based program for patients with the highest complexity. Non-hospital-based sleep laboratories may need to have different algorithms. A successful safety protocol will need utilization of available resources, such as an RR team or its equivalent. Free-standing centers may have no option but to trigger a 911 call.

False alarms are inevitable, and expensive, but patient safety is the key endpoint. False alarms, or unnecessary activation of the RR team, were rare in our center. Of the 40 incidents that resulted in activation of the RR team, 30 patients were transferred to the ED. Of the 10 patients who completed the study without transfer to the ED, 2 required intervention for hypoglycemia and 2 were given supplemental oxygen (temporary oxygen given to 1 patient following a seizure consistent with typical break-through episodes of refractory epilepsy as confirmed by a family member). The remaining 6 patients were evaluated at bedside without further intervention (6/40, 15%), although even in these patients, activation of the RR team was likely reasonable: for example, tachycardia in the setting of not taking prescribed β-blocker, or mild hypoxia following a break-through seizure before return to baseline neurologic status as confirmed by a family member. Avoidance of false alarms is critical in any system of evaluation, and knowledge of medical history can reduce unnecessary activations. For example, adequate knowledge of baseline cardiac status will minimize calling an emergency based on detection of apparently new atrial fibrillation or nonsustained ventricular tachycardia, and knowledge of complex partial or focal motor seizure history can alleviate technician discomfort with break-through events if these are reasonably frequent for a patient.

There are some limitations in our report. The first is that our center does not test children, in whom unique challenges for sick patients may be seen and need modified protocols—conditions such as neuromuscular disease and respiratory failure, epilepsy, and severe behavioral outbursts will require thoughtful interventions. Second, our center uses a hospital-based sleep laboratory for the majority of our in-laboratory studies, where much of the infrastructure for an RR to medically acute situations is in place. Given the limited number of patients in our affiliated free-standing sleep laboratories, this study does not fully address protocols for those facilities. However, the protocols for acute symptom assessment, sedative use policy, and tracheostomy procedures can be used in free-standing sleep laboratories.

Risk stratification can reasonably occur at multiple levels: ordering and scheduling, prestudy screening and vital sign measurement, and vigilance once the test has started. As routine diagnostic testing localizes primarily to the home environment, and improved sophistication of positive-pressure devices enables bypass of the traditional sleep laboratory, selection for more high-risk patients with increased medical comorbidities in the sleep laboratory is inevitable. Standardized triage and response protocols can optimize recognition and reaction to emergencies and maximize patient safety in the sleep laboratory.

DISCLOSURE STATEMENT

All authors have reviewed and approved this manuscript. Dr. Thomas discloses the following: (1) licensed patent (ECG-spectrogram) and royalties through Beth Israel Deaconess Medical Center to MyCardio, LLC; (2) an unlicensed patent for a CO₂ device to treat central/complex apnea; (3) licensed patent and royalties for an auto-continuous positive airway pressure algorithm through Beth Israel Deaconess Medical Center to DeVilbiss-Drive; (4) consulting for Jazz Pharmaceuticals, Guidepoint Global, and GLG Councils. The other authors report no conflicts of interest.

ABBREVIATIONS

ED

emergency department

RR

rapid response