Increase in tracheostomy for prolonged mechanical ventilation in North Carolina, 1993-2002

Objective: Patients who require tracheostomy for prolonged mechanical ventilation have poor outcomes and high costs of care. However, recent longitudinal trends relevant to these patients and their care have not been described. We aimed to describe trends in the annual incidence and timing of tracheostomy for prolonged mechanical ventilation, as well as prolonged mechanical ventilation patient resource utilization and overall in-hospital mortality. Design and Setting: Retrospective review of the North Carolina Hospital Discharge Database, a comprehensive record of all state nonfederal, nonpsychiatric hospital discharges between 1993 and 2002. Patients: Patients were 9,794 medical and surgical patients greater than or equal to18 yrs of age with International Classification of Diseases, Ninth Revision, Clinical Modification code 96.72 (mechanical ventilation for >96 hrs) and Diagnosis Related Group code 483 (tracheostomy except for face, neck, and mouth diagnoses). Interventions: None. Measurements: Incidence rates adjusted for annual population growth, mechanical ventilation days until tracheostomy placement, length of stay, and hospital charges and payments adjusted by the medical component of the Consumer Price Index. Main results: Between 1993 and 2002, the incidence of tracheostomy for prolonged mechanical ventilation increased across all age groups from 8.3 of 100,000 to 24.2 of 100,000 (p < .001), although most significantly among patients <55 yrs of age. During this period, a decrease was seen in mortality (from 39% to 25%), median mechanical ventilation days to tracheostomy placement (from 12 to 10 days), and median length of stay (from 47 to 33 days). By 2002, patients were almost three times less likely to be discharged to home independently although twice as likely to be sent to a skilled nursing facility. Although prolonged mechanical ventilation patients with tracheostomies represented only 7% of all who required mechanical ventilation, their total charges during the study period were $1.74 billion-22% of all mechanical ventilation patient charges. Conclusion: The incidence of tracheostomy for prolonged mechanical ventilation increased by nearly 200% during the past decade in North Carolina, exceeding changes in the overall incidence of respiratory failure three-fold. Although in-hospital mortality, length of stay, and charges per patient fell over time, the overall resource utilization of prolonged mechanical ventilation patients increased dramatically.