Methods

This project was retrospective in design with 2 phases of review. The first phase comprised data collection 6 weeks before the addition of the ESS in the EMR (time 1) and 6 weeks after the ESS was added to the EMR (time 2). The second phase of the study included staff education on the use and interpretation of the ESS, followed by data collection 6 weeks after implementation of the ESS for all patients in the internal medicine office (time 3). Records were reviewed in a random and systematic fashion to obtain the sample most representative of patients seen on a daily basis. Participants’ approval was obtained from a university institutional review board, and the study was determined to pose minimal risk to participants.

Population and Setting

Men and women between the ages of 18 and 92 years were included, and 50 participants were identified from each 6-week period, totaling 150 patients. A list of patients was obtained from the EMR, and every fifth patient was selected until 50 participants from each time frame were selected.

The project was conducted in an internal medicine office setting with 2 medical providers and an average patient load each of 10 to 20 patients per day. Office staff were responsible for preparing the patient for the office visit and examination. Staff were educated on the ESS and how to quickly and accurately complete the assessment with each patient.

Instrument

The ESS is a commonly used tool to determine daytime sleepiness and is useful in alerting providers of a potential sleep disorder such as OSA.6,13 The ESS consists of 8 standard questions that can be quickly administered. Each question is framed to elicit the likelihood of dozing during different common daily activities. A score of 0 is given when the patient reports that they would never doze during an activity, 1 for a slight chance of dozing, 2 for a moderate chance, and 3 for a high chance.15 The ESS score is the sum of all answers and can range from 0 to 24. Scores less than 10 indicate a normal level of daytime sleepiness. A score of 10 or above is considered abnormal, and the patient should be referred for further testing.13 According to Johns, the ESS has a reliability of 0.81; extremely high sensitivity and specificity scores of 93.5% and 100%, respectively; and major advantages over other tests of daytime sleepiness.16

Demographic data including age, sex, height, weight, body mass index (BMI), race, and insurance were collected. The ESS numeric score was recorded and documentation was made if the patient was referred for polysomnography.

Procedure

Phase I. Medical records of patients seen in the 6 weeks before adding the ESS to the EMR in 2016 were obtained. These charts were reviewed and marked as either having an ESS score calculated during the visit or an ESS score not obtained. Those with ESS scores were recorded with the numeric score, scoring of 10 or more, and documentation of referral for polysomnography was made. A second set of data was collected in the same way from the medical records of patients 6 weeks after the ESS was added to the EMR in 2016.

Phase II. Staff were educated on the ESS tool and how to properly obtain and document the ESS in the EMR, and they were instructed to notify the provider if the ESS score is 10 or greater. Handouts were provided to all staff who may be rooming patients on the ESS, and contact information was provided for questions. One week of learning was provided for the staff to adjust to performing an ESS on each patient. A second meeting was held with staff for questions, comments, and possible changes. The ESS was then used during every patient encounter in the office. Weekly check-ins with staff using and documenting the ESS were conducted during the implementation phase. Each week during the implementation phase, 10 medical records were randomly selected to ascertain whether the ESS was documented correctly. When inconsistencies were noted, staff retraining and coaching were provided.

Six weeks after implementation of ESS screening of all patients in the medical office, a third set of 50 medical records of patients was identified and records were reviewed for every fifth medical record for patients seen in the prior 6-week period, until a total of 50 patients were identified.

Data Analysis

Demographic characteristics and clinical variables were analyzed using descriptive statistics. Means, standard deviations, and ranges were calculated for continuous variables at all 3 periods, including age, sex, race, insurance, and BMI. Sleep characteristics were analyzed using mean and standard deviations for ESS scores. After collection of the ESS scores, data were converted to either normal (ESS <10) or abnormal (ESS ≥10), and documentation was made if the patient was referred for polysomnography.

Results

Demographic characteristics of participants are displayed in Table 1. Patients at time 1 were slightly older than patients at the other 2 periods. There were nearly equal percentages of male and female patients sampled over the course of all 3 periods, with the exception being that time 3 has slightly more female patients. The sample consisted of predominantly white individuals, with Hispanics or Latinos and African Americans making up the remainder of patients sampled. The majority of patients in time 1 and time 2 had private insurance and/or Medicare; however, in time 3, 100% of patients were enrolled in a fee-based care model. Mean BMI was similar across the 3 periods.

Table 1. Study Demographic and Clinical Characteristics.

Characteristic Time 1 (n=50) Time 2 (n=50) Time 3 (n=50)
Age, mean/SD (range) 71±13 (19-88) 63±16 (24-88) 64±16 (22-93)
Sex, (n/%)      
  Male 25 (50%) 25 (50%) 17 (34%)
  Female 25 (50%) 25 (50%) 33 (66%)
Race, (n/%)      
  White 49 (98%) 44 (88%) 48 (96%)
  African American  0 (0%)  2 (4%)  0 (0%)
  Hispanic or Latino  1 (2%)  4 (8%)  2 (4%)
Insurance, (n/%)      
  Private 16 (32%) 26 (52%)  0 (0%)
  Medicare 34 (68%) 24 (48%)  0 (0%)
  Fee-based care  0 (0%)  0 (0%) 50 (100%)
BMI, mean/SD/range 29 ± 6 (17-42); n=49 30 ± 6 (18-46); n=46 28 ± 3 (26-32); n=4

Data on sleep characteristics revealed no ESS scores documented in the EMR at times 1 and 2, despite the availability of the ESS in the EMR at time 2 (Table 2). At time 3, with screening of all patients, 10 patients had an ESS of 10 or more and were referred for polysomnography.

Table 2. Study Sleep Characteristics

Characteristic Time 1 (n=50) Time 2 (n=50) Time 3 (n=50)
ESS score: mean/SD 0±0 0±0  6±4
ESS score ≥ 10 (n) 0 0 10

Discussion

The key findings from this quality improvement project illustrate that daytime sleepiness, use of the ESS as a common screening tool, and referrals for polysomnography are missed opportunities to identify patients with a potential diagnosis of a highly treatable condition such as OSA. Approximately 80% of individuals who report poor sleep have undiagnosed and untreated OSA, indicating an imminent need for routine screening.17 At this time, there is limited discussion in the literature about screening every patient for OSA, regardless of risk factors. Although numerous sources cite the usefulness of the ESS as a screening tool in individuals who are suspected of having OSA, there is limited research on screening asymptomatic adults.

Garbarino et al reported that screening dangerous-goods truck drivers for OSA revealed an unexpected high prevalence of sleep apnea in this population.18 The study analyzed 283 truck drivers who did not have symptoms of OSA. After screening was implemented, almost half of asymptomatic individuals were flagged as high risk. These high-risk individuals underwent polysomnography, and 35.7% of the group was confirmed with a diagnosis of OSA.18 These individuals were required to begin continuous positive airway pressure therapy, and after 2 years, the rates of motor vehicle accidents and near-miss accidents were significantly reduced.18

The US Preventive Services Task Force has issued a call for further research on screening asymptomatic adults for OSA through the use of screening questionnaires such as the ESS. At this time, the task force states that there is a critical gap in information and insufficient data to recommend routine screening. The task force also reports that only 20% of patients with sleep-related symptoms spontaneously reported their symptoms to their primary care provider; this leaves 80% of individuals who have sleep-related symptoms neither reported to their provider nor adequately assessed for OSA. Routine screening for OSA could potentially capture this remaining 80% and provide the necessary testing and treatment that is so desperately needed. Nurse practitioners and physician assistants can play a key role in integrating screening for OSA in practice.19

The ESS is a simple, quick, and easily administered screening tool that can be used in primary care, specialty practice, and hospital settings. During this quality improvement project, office staff asked each patient the screening questions while obtaining the patient’s vital signs, which required little or no extra time with the patient. Taking a few moments to ask each patient 8 simple questions has the potential to decrease risk and safety issues related to OSA, as well as to improve patient outcomes and quality of life.