Hospital screening for obstructive sleep apnea in patients admitted to a rural, tertiary care academic hospital with heart failure.

Abstract

Background:

Rural communities represent a vulnerable population that would significantly benefit from hospital based OSA screening given these areas tend to have significant healthcare disparities and poor health outcomes. Although inpatient screening has been studied at urban hospitals, no study to date has assessed this approach in rural populations.

Methods:

This study utilized the Electronic Medical Record (EMR) to generate list of potential candidates by employing inclusion/exclusion criteria as screening. Subjects identified were then approached and offered information regarding the study. Screening for OSA entailed a tiered approach utilizing the sleep apnea clinical score (SAC) and portable sleep testing. Individuals identified as high risk (SAC ≥ 15) for OSA underwent evaluation with a portable sleep testing system while hospitalized. All participants with an apnea-hypopnea index (AHI) ≥ 5 events/h confirmed by a sleep medicine physician were considered screen positive for OSA. If approved/available, subjects screening positive for OSA were provided with an auto-titrating continuous positive airway pressure (PAP). Patient characteristics were analyzed using descriptive statistics. Categorical data were described using contingency tables, including counts and percentages. Continuously scaled measures were summarized by median with range. This study was registered with ClinicalTrials.gov. Identifier: NCT03056443.

Results:

Nine hundred and fifty eight potential subjects were identified. The three most common reasons for exclusion included previous OSA diagnosis or exposure to PAP therapy (n=357), advanced illness (n=380) and declined participation by the individual (n=68). The remaining 31 subjects underwent further evaluation for obstructive sleep apnea. Twenty three subjects had a high sleep apnea clinic score. Per our study protocol, 13 subjects who screened positive for OSA were initiated on APAP therapy.

Conclusions:

Our study provides important insight into the burden of sleep disordered breathing (SDB) and unique challenges of hospital based OSA screening/treatment in a rural setting. Our study identified barriers to successful screening in a rural population that may be well addressed by adapting previous research in hospital sleep medicine.

Introduction:

In recent years, there has been growing interest regarding screening for OSA in hospitalized patients [1]. Research evaluating portable sleep monitoring devices and high-resolution pulse oximetry (HRPO) suggest they can be utilized in screening hospitalized patients for obstructive sleep apnea [2, 3]. Limited data to date suggests that inpatient screening of OSA may improve outcomes; particularly in specific groups such as those with heart failure, acute stroke, COPD and obesity hypoventilation syndrome [4–8]. While this data is promising, the majority of studies evaluating hospital based sleep apnea screening have been completed at large urban centers [4, 9]. Hospital based OSA screening programs may be of particular importance in rural centers given the limited access and other healthcare disparities in these populations.

Obstructive sleep apnea (OSA) syndrome is a common clinical problem that is underdiagnosed and undertreated even in modern, urban settings [10]. OSA is associated with multiple adverse health consequences including cardiovascular disease [11]. The prevalence of OSA in adults had been estimated to be 2–4% [12]; however, more recent data suggests that up to 14% of men and 5% of women have OSA [13]. This rise in prevalence has been attributed in part to the obesity epidemic, and no region has been hit harder by the obesity epidemic than rural America and Appalachia. According to the CDC’s Behavioral Risk Factor Surveillance System (BRFSS), West Virginia has the highest obesity rate of any state in the U.S. at 37.7% [14]. Given the region’s high prevalence of obesity (a known risk factor for OSA) as well as leading the nation in respiratory and cardiac disease, improved identification and treatment of obstructive sleep apnea would have a positive impact on the health of this underserved rural population [15–21].

West Virginia has the highest obesity prevalence in the U.S as well as one of the highest rates for cardiovascular disease in the country, thus improved identification and treatment of obstructive sleep apnea could have substantial health benefits to the community. Despite federal support and other programs aimed at improving care for rural communities, these populations have failed to overcome their healthcare challenges. Our local tertiary care university hospital is unique, because the population we serve is almost exclusively rural dwelling [22]. Thus, our institution is well suited to evaluate novel hospital care delivery models for health centers serving rural populations. To better serve this under-represented population, we developed a pilot project to assess the feasibility of hospital screening for obstructive sleep apnea in a rural hospital and gain a better understanding of the disease burden in our local Appalachian community.

Methods:

Study subjects were identified using a convenient and established heart failure admission tracking system, which used ICD-9 codes (changed to ICD-10 codes in 2016) embedded in the electronic medical record (EMR). This heart failure admission tracking system was part of an institutional quality initiative as it is the most common cause of admission to the hospital in the United States. New admissions were screened in the EMR on a daily basis by the study team. Both inclusion and exclusion criteria were first assessed through review of the EMR.

Inclusion criteria were:

1. Age ≥18 years old.

2. Able to provide written informed consent (including agreement to privacy language within the informed consent or in ancillary documents compliant with HIPPA before the initiation of any study–related procedures)

3. Hospitalized patients having a documented history of congestive heart failure in agreement with the 2013 ACCF/AHA heart failure definition [23]

4. Anticipated hospitalization of more than 24 hours.

5. Willing to comply with the protocol.

Exclusion criteria were:

1. Established obstructive sleep apnea and/or previous exposure to positive airway pressure (PAP) therapy.

2. The presence of any active conditions that the investigators felt would interfere with testing or potential therapy (hemodynamic instability, respiratory failure, unconsciousness, pneumothorax, penetrating chest trauma, persistent nausea/vomiting, facial anomalies/facial trauma, upper gastrointestinal bleed or history of recent gastric surgery).

3. Presence of a clinically significant illness or medical condition that the investigators felt would prohibit the subject from participating in the study (a number of subjects were ruled due to advanced COPD, malignancy or having a high baseline home oxygen requirement).

4. Declined testing or potential therapy.

5. Patient not well enough to return to home environment, (i.e. transferred to rehab, hospice or skilled nursing facilities, etc).

Individuals were screened during the weekdays based on our electronic medical record generated list. The medical record was reviewed for inclusion and exclusion criteria. Subjects identified were approached and offered information regarding the study. They were further assessed regarding inclusion and exclusion criteria. Individuals were offered enrollment if eligible and interested in participating. Subjects were provided written informed consent before tiered screening was initiated. Once enrolled, baseline demographic data, NYHA class, comorbid illnesses (diabetes, hypertension, hyperlipidemia, coronary artery disease, atrial fibrillation, depression and stroke), echocardiogram results and zip code (as a rough estimation of distance from our sleep center) were recorded.

Subject’s risk for obstructive sleep apnea was then assessed utilizing the sleep apnea clinical score (SAC) [24]. Individuals identified as high risk (SAC ≥ 15) for OSA were asked to complete the Epworth Sleepiness Scale and Minnesota Living with Heart Failure Questionnaire [25, 26]. The following evening, hospitalized subjects with a SAC ≥ 15 underwent evaluation with an FDA approved level III portable sleep testing system (Resmed ApneaLink™) [27]. All participants with an apnea-hypopnea index (AHI) ≥ 5 events/hour that was confirmed by a sleep medicine physician were considered screen positive for OSA. If approved/available, subjects screening positive for OSA were provided with an auto-titrating continuous positive airway pressure (APAP) device through insurance or provided with a loaner CPAP device through our sleep center. All subjects initiated on therapy were placed in an auto-titrating mode. Individuals who demonstrated central or complex sleep disordered breathing (SDB) that would not be appropriately addressed by auto-CPAP based on review by a sleep medicine specialist were referred to the sleep center for evaluation.

Patient characteristics were analyzed using descriptive statistics. Categorical data were described using contingency tables, including counts and percentages. Continuously scaled measures were summarized by median with range. This research was supported by NIH/NIGMS award number and approved by the West Virginia University IRB (Approved IRB protocol number 150982524). This study was registered with ClinicalTrials.gov. Identifier: NCT03056443. This research was conducted following the principles of the Declaration of Helsinki.

Results:

Nine hundred and fifty-eight individuals were screened in the electronic medical record. Of these, 927 met exclusion criteria. Figure 1 shows reasons for exclusion. The three most common reasons for exclusion included previous OSA diagnosis or exposure to positive airway pressure therapy (n=357), advanced illness (n=380) and declined participation by the individual (n=68). The remaining 31 subjects underwent further evaluation for obstructive sleep apnea.

Figure 1.

Tiered approach for obstructive sleep Apnea screening.

Following our tiered screening approach of this study group, 13 subjects screened positive for obstructive sleep apnea (SAC ≥ 15 and AHI ≥ 5). Eighteen subjects did not show obstructive sleep apnea. A number of individuals did not complete portable sleep testing due to a low sleep apnea clinical score, being discharged prior to this evaluation or having the sleep study removed prior to adequate data being obtained. Two subjects were found to not have obstructive sleep apnea but complex apnea (Cheynne-Stokes Respirations) and were referred immediately to the sleep center due to the significance of this finding. Of the 23 subjects with a high sleep apnea clinic score, only one had a negative screening portable sleep study.

Per our study protocol, subjects who screened positive for OSA based on our tiered approach were discharged home with APAP therapy. Baseline characteristics of this population can be found in table 1. Our tiered approach did identify patients with significant symptoms (average Epworth Sleepiness Scale of 15.4) and on average moderate severity of obstructive sleep apnea (average apnea hypopnea index of 16.4). Subjects were asked to follow-up at 1 month and six months as is our standard clinical practice following PAP initiation for OSA. Standard clinical follow-up information was obtained including APAP download and the Epworth Sleepiness Score. All subjects were initiated on auto-titrating CPAP therapy. To support follow-up subjects were provided with $20 gift cards for travel costs. Of the thirteen subjects initiated on treatment at discharge, only 3 (23%) were adherent to therapy at six months (figure2).

Table 1:

Baseline Characteristics of Subjects Screening Positive for Obstructive Sleep Apnea

Baseline Characteristics (units/definition)	Average (Range or Number)
Age (years)	57.9 (35 – 75)
Epworth Sleepiness Scale Score	15.4 (5 – 24)
Minnesota Heart Failure Questionnaire Score	71.7 (9 – 100)
Distance Lived from Sleep Center (miles)	55.7 (6 – 183)
Apnea Hypopnea Index (events/hour)	16.4 (5.6 – 56.7)
Gender	38.4% female (5)
Systolic Heart Failure (Ejection Fraction <40%)	46.1%
New York Heart Association Functional Class	II – III
Race	92% Caucasian (12) 8% African American (1)

Figure 2.

Bar graph reporting subjective adherence to positive airway pressure therapy at two weeks objective adherence to positive airway pressure therapy at one month and six months from machine download.

Discussion:

Our study provides important insight into the burden of sleep disordered breathing and unique challenges of disease identification and effective treatment in a rural setting. Despite a heavy burden of sleep disordered breathing, the current screening protocol captured a limited number of patients. Furthermore, the intervention program suggested sub-optimal outcome in terms of adherence to therapy, which is not consistent with prior publications [4,27]. These results suggest our rural population may have variables impacting the outcome over and beyond those described in urban settings. Understanding these variables may be of particular importance in institutions serving rural communities and establishing successful hospital OSA screening programs. We hypothesize that some of these variables may include lower literacy rates, limited access to durable medical equipment (DME) support and lower socioeconomic status which has been suggested in other studies [28,29].

Of the 958 individuals identified through the electronic medical record, nearly three quarters of subjects were excluded from evaluation due to previous OSA diagnosis/PAP exposure or advanced illness. Thus a large number of individuals had prior assessment for OSA or were previously treated with positive airway pressure therapy/non-invasive ventilation. However it is likely a significant number of these patients were not on treatment for OSA or ineffectively treated [30–32]. Our research team had concerns that individuals with advanced comorbid disease would not be suitable candidates for hospital based OSA screening. For instance many patients were excluded due to significant comorbid COPD identified on EMR review. In the past, data on prevalence of OSA in COPD has been equivocal [33, 34]. However, recent work, specifically in hospitalized COPD patients show almost 50% prevalence of OSA impacting both hospital readmission and mortality adversely [35, 36].

Our tiered approach did identify patients with significant symptoms (average Epworth Sleepiness Scale of 15.4) and on average moderate severity of obstructive sleep apnea (average apnea hypopnea index of 16.4), both of which may have significant health consequences [37,38]. Although only a small number of subjects were initiated on therapy, our study is the first to provide insight into the large burden of sleep disordered breathing in hospitalized patients from a rural setting. Furthermore, this study highlights potential barriers to screening in patients hospitalized specifically from rural settings (figure 2). This data serves as an important baseline for further refinement of hospital based OSA screening models designed to impact health outcomes in institutions serving rural communities. We believe, with further development, hospital based OSA screening programs implemented at rural centers will lead to improved patient outcomes as shown in studies done at urban settings [7–9].

Our study has several important findings, but it should be interpreted with the following limitations. First, due to logistical limitations, the screening was limited to an established heart failure admission tracking system embedded in our EMR. Thus, only select medical services caring for heart failure patients were included and our findings cannot be generalized to other services. Second, the screening was limited to weekdays, but this protocol was consistent with prior studies [21,39]. Finally, we recognize that a number of subjects were likely missed as the EMR algorithm that was aimed at identifying heart failure patients and not specific for sleep apnea screenings.

Our results suggest a high prevalence of OSA in our local population. To better serve this unmet need, we have developed a robust hospital sleep medicine program for our rural academic institution. Our study identified barriers to successful screening in our local rural population that we believe can be overcome by previous research in hospital sleep medicine. For instance, use of high-resolution pulse oximetry (HRPO) has recently been shown to be effective for OSA screening in hospitalized patients. This technology provides a relatively simple, low-cost, and comfortable alternative to the level III portable sleep tests utilized in this hospital study [40, 41]. Our tiered screening approach was found to be very specific (only one negative sleep study) but was likely inadequately sensitive to effectively screen our population. Thus, our new tiered screening program utilizes the EMR to generate a specific screening list for OSA based on BMI and includes an increased number of services. Our new tiered screening program also uses a validated screening questionnaire (STOP Questionnaire) that is simpler for support staff to administer compared to the SAC and has been successfully utilized in other studies for hospital based OSA screening [21, 24, 42]. Finally we hope to ensure identified patients are appropriately diagnosed and treated for OSA in the outpatient settings, which we had little success of accomplishing during this pilot screening trial.

Based on our experience and the results of this study we have instituted a revised protocol that addresses the variables affecting better implementation and outcome of this screening program. The issues addressed include screening all high risk patients regardless of comorbid disease, not excluding patients with a reported history of SDB or previous therapy exposure and a robust educational program including an accommodation trial of PAP therapy while hospitalized with a dedicated and well trained respiratory therapist to improve outpatient follow-up and treatment adherence (figure 3) [43–48].

Figure 3.

Hospital based OSA screening algorithm for clinical hospital sleep medicine service.

Conclusion

In conclusion, our study provides important insight and baseline data into the burden of sleep disordered breathing, the unique challenges of disease identification, and effective treatment in patients hospitalized in a rural setting. Our results suggest a high disease prevalence in our local rural community. Hospital screening for obstructive sleep apnea may represent an important way healthcare centers can address this significant healthcare disparity in rural populations. Our study identified barriers to successful hospital screening of SDB in an institution serving predominantly rural community. We believe these barriers will be well addressed by our modified screening protocol (figure 3). This baseline data will help determine if targeted interventions as mentioned above will help improve hospital screening and outcomes of SDB in this cohort with significant healthcare disparities.