First Large-Scale Study Leverages Big Data for Small Sleepers

Infants and Young Children with SDB an Understudied Minority

Though sleep disorders are common in the pediatric population, infants represent an understudied minority. Little research has been published on the subject of sleep-disordered breathing (SDB) in this population, even though it can have a significant impact on health during the formative years of neurocognitive development.

To address this issue, Zarmina Ehsan, MD, Pediatric Sleep Medicine, Children’s Mercy Kansas City, joined forces with the data science team led by Mark Hoffman, PhD, Chief Research Information Officer with the Children’s Mercy Research Institute, home to one of the most progressive and dynamic pediatric research programs in the world.

They utilized big data to conduct the first large-scale study on SDB in infants. Dr. Ehsan identified the gaps in currently available literature as:

1. Limited understanding of the evolution of SDB during infancy and beyond.
2. Paucity of centers performing infant polysomnography (PSG) and, therefore, available literature generated based on information from select centers.
3. Management of infant obstructive sleep apnea (OSA) based on extrapolation of available literature from older children, potentially increasing the risk of providing inappropriate therapy to this understudied and underserved population with SDB.

These gaps also make this population an ideal focus of data science.

Database Highlights

Children’s Mercy utilizes the de-identified electronic health record Cerner Health Facts. The hospital also manages the Standardized Health and Research data Exchange (SHaRE), a group of Health Facts contributors who have agreed to collaborate and share de-identified patient data. By employing this big data methodology, the researchers determined the demographics, sleep diagnoses, comorbid medication conditions, healthcare utilization and economic outcomes for this patient database. Earl Glynn, a data scientist in Dr. Hoffman’s group, developed an innovative method to characterize comorbidities in this data resource.

Highlights included:

• In a cohort of 68.7 million unique patients, over a nine-year period, there were 9,773 infants and young children with a diagnosis of SDB (obstructive sleep apnea [OSA], nonobstructive sleep apnea, and “other” sleep apnea) who met inclusion criteria, encompassing 17,574 encounters, and a total of 27,290 diagnoses across 62 U.S. health systems, 172 facilities and three patient encounter types (inpatient, clinic and outpatient).
• Thirty-nine percent of patients were female; 39% were 1 year of age (6,429); 50% were 1-2 years of age; and 11% were 2 years of age.
• The most common comorbid diagnoses were micrognathia, congenital airway abnormalities, gastroesophageal reflux, chronic tonsillitis/adenoiditis and anomalies of the respiratory system.
• Payer mix was dominated by government-funded entities.

Big Data Yields Insight into Comorbidities

This is the first study to utilize big data in this understudied population and to demonstrate the utility of large-scale de-identified electronic health record data in sleep research in general.

Interestingly, this study confirmed what we see clinically: that sleep-disordered breathing in infants and young children is multifactorial and closely linked to comorbid medical conditions.

Additional insights that are not widely known include:

• Chronic tonsillitis and adenoiditis were prevalent in 21% of the OSA cohort, compared to 0.6% of the base population within this age group (prevalence ratio of 36%).
• Adenotonsillar hypertrophy is a well-known risk factor for OSA in older children, although few studies also report that it can be prevalent as early as infancy.
• About 2.9% of this cohort had a diagnosis of jaw abnormality, compared to 0.1% in the general population within this age group. This phenotype had the second largest prevalence ratio (27.9), but was not present in the nonobstructive sleep apnea or “other” sleep apnea cohorts, which is not unexpected as micrognathia is a known risk factor for OSA in infants, particularly in those with a cleft palate.
• Congenital anomalies of the respiratory system were prevalent in 11% of the OSA cohort, compared to 0.8% of the base population. Retrospective clinical research has shown that congenital soft tissue airway abnormalities or bone abnormalities involving the face also predispose to OSA.
• Laryngomalacia is one of the leading causes of stridor, upper airway obstruction, and therefore OSA, in infants. Other airway abnormalities such as subglottic stenosis or tracheomalacia are often present as well.
• Prevalence ratios for “dependence on respirator/ventilator” and “other diseases of the lung” also ranked high at 11% each.
• A diagnosis of otitis media was also prevalent within the cohort. There are a handful of studies suggesting a high prevalence of middle ear effusion and otitis media in infants with OSA (particularly those with Down syndrome).
• Chromosomal abnormalities consistently scored high on prevalence ratio in all three cohorts, with a higher prevalence than in the base population.
• The link between gastroesophageal reflux and OSA in infants has been reported in the literature; however, the evidence for this is weak. In all three cohorts, this research found a higher prevalence of gastroesophageal reflux compared to the base population, but prevalence ratios did not appear to be significant.

Conclusions:

Dr. Ehsan and her colleagues concluded large-scale aggregate, de-identified EHR data provide a rich and untapped resource to examine these conditions. The team is in the process of pursuing further research targeting infants and young children to better understand the long-term outcomes of SDB in this patient population.¹

Learn more about Pulmonary and Sleep Telemedicine Services at Children’s Mercy

Zarmina Ehsan, MD, Pediatric Pulmonology, Pediatric Sleep Medicine; Clinical Assistant Professor of Pediatrics, University of Missouri-Kansas City School of Medicine

zehsan@cmh.edu • (816) 234-3700

For consults, admissions or transport call: 1 (800) GO MERCY / 1 (800) 466-3729.

References:

1. Ehsan Z, Glynn EF, Hoffman MA, Ingram DG, Al-Shawwa B. Small Sleepers, Big Data: Leveraging Big Data to Explore Sleep-Disordered Breathing in Infants and Young Children. SLEEPJ, 2020, 1–13. doi: 10.1093/sleep/zsaa176.