Early detection of hospitalized patients with previously diagnosed obstructive sleep apnea using computer decision support alerts.

Obstructive sleep apnea (OSA) is a worldwide problem affecting 2-14% of the general population and most patients remain undiagnosed. OSA patients are at elevated risk for hypoxemia, cardiac arrhythmias, cardiorespiratory arrest, hypoxic encephalopathy, stroke and death during hospitalization. Clinical screening questionnaires are used to identify hospitalized patients with OSA; especially before surgery. However, current screening questionnaires miss a significant number of patients and require more definitive testing before specific therapy can be started. Moreover, many patients are admitted to the hospital with a previous diagnosis of OSA that is not reported. Thus, many patients with OSA do not receive appropriate therapy during hospitalization due to the lack of information from previous inpatient and outpatient encounters. Large enterprise data warehouses provide the ability to monitor patient encounters over wide geographical areas. This study found that previously diagnosed OSA is highly prevalent and undertreated in hospitalized patients and the use of early computer alerts by respiratory therapists resulted in significantly more OSA patients receiving appropriate medical care (P < 0.002) which resulted in significantly fewer experiencing hypoxemia (P < 0.006). The impact was greater for non-surgery patients compared to surgery patients.