Tracheostomy is a medical procedure performed on the front of a persons neck. It is used to create a connection between the persons trachea and a mechanical ventilator instead of using a tube going through the mouth into the trachea, oral intubation. Living with a tracheostomy tube is less stressful compared to oral intubation and facilitate being awake and the start of training on spontaneous ventilation in mechanically ventilated patients. Studies of the timing of tracheostomy are either severely affected by methodological bias of to small to determine an effect. Thus, it is not known what the optimal timing of the tracheostomy is in mechanically ventilated COVID-19 patients.

The study hypothesis is that a strategy of tracheostomy during the second week of mechanical ventilation yields more IMV-free days and lower mortality than continued mechanical ventilation without tracheostomy, when efforts are made to neutralize immortal time bias.

The hypothesis has been slightly changed to accommodate the methods change described below.

Data sources Existing data provided for another project will be used.

Statistical methods:

A Markov multistate model with inverse probability weighting and landmark analyses at 7, 14, and 21 days from IMV start. The primary outcome is estimated ventilator-free days alive at day 60; secondary analyses includes 60-day mortality.

The methods has been changed to multistate + landmark because the data did not support a full cloning, censoring, weighing approach.

Inclusion Criteria:

• Adult
• admitted to an intensive care unit in Sweden from 2020 to 2021
• Main discharge diagnosis COVID-19 (ICD-10, U07.1)
• Invasive mechanical ventilation

Exclusion Criteria:

• No Swedish personal identification number.

Invasively mechanically ventilated adults in the Swedish intensive care registry.

Individual patient data will be shared under the regulations of the European general data protection act and after researchers obtain relevant permissions from the Swedish ethical review authority.