Using a joint continuous time Markov chain to represent the trauma patient's ICU and ventilator experience
Patients who require prolonged mechanical ventilation (MV) are highly prone to a variety of complications, including but not limited to pulmonary edema, ventilator-associated pneumonia, erosive esophagitis, or death. The ability to accurately and reliably predict the number of MV days a patient will have upon entry to the emergency department (ED) would be instrumental in helping clinicians determine whether the patient would benefit from preventive procedures, such as an early tracheostomy, to prevent MV complications and reduce the patient's number of days requiring MV and intensive care unit (ICU) use. Traditionally, investigators use a composite outcome to calculate a trauma patient's ventilator-free, ICU-free, and hospital-free days. Although the use of a composite outcome is the standard approach in trauma clinical trials, we believe that vital information about patients' ventilator and ICU use is ignored when the patient dies before the end of the study period or remains on the ventilator or is in the ICU the entire study period. We believe the composite outcome does not give doctors an accurate picture of a patient's total ventilator and ICU experience. We have developed a joint continuous-time Markov chain (CTMC), an analytical prediction model, and estimation technique that allow for a more accurate estimate of an individual patient's predicted number of ventilator and ICU days based on the patient's unique baseline characteristics collected at ED arrival, such as age and mechanism of injury. This methodology produces an estimation of ventilator days that is within one day of the true MV duration. By using our methodology to attain an accurate estimate of a particular patient's predicted number of MV days, a clinician possesses a new piece of critical information to determine whether a patient would benefit from an early tracheostomy.^
Seay, Roann, "Using a joint continuous time Markov chain to represent the trauma patient's ICU and ventilator experience" (2015). Texas Medical Center Dissertations (via ProQuest). AAI3720107.