Data collection

A list of patients with tracheostomy placement during a hospital admission was extracted from the hospital administrative database. Records that contained International Statistical Classification of Diseases and Related Health Problems 10th Revision version 9 – Australian modified version (ICD10AM), procedure codes for percutaneous tracheostomy (4188000), surgical tracheostomy, temporary (4188100) and surgical tracheostomy, and permanent (4188101) were extracted for further analysis. Each recorded procedure was manually checked and the following information fields were extracted – unit record number, age, gender, number of hours on invasive mechanical ventilation, day of admission to ICU, ICD10 diagnoses, ICD10 procedures, length of stay (LOS) in ICU, LOS in hospital, and hospital mortality. Furthermore, data for each patient were manually cross-checked from the surgical database and the date of tracheostomy was recorded. These records were linked by a unique hospital unit record number to a hospital costing database, where the costs of the care pathway for each patient in hospital is collated and recorded. We examined the hospital costing database to identify tracheostomy patients as those having laryngectomy or permanent tracheostomy as a part of primary cancer/ neck surgeries, and those who underwent tracheostomy as a part of their ICU care. ICU patients undergoing tracheostomy were further examined by procedure: percutaneous procedure (at the bedside) versus surgical tracheostomy (in the operation theatre); and timing of procedure: early (≤10 days) and late (>10 days) after ICU admission. We chose a cut-off point of 10 days to define early and late tracheostomy to achieve consistency with the definitions adopted by a recent review article.¹⁷