Early prediction and prevention of prolonged weaning in high risk patients via identification through a computer model and additional assistance in treatment through certified weaning centers in Baden-Württemberg

Background: Invasive mechanical ventilation has been a standard therapy for patients undergoing intensive care, and has gained in importance especially due to the Covid-19 pandemic. With increasing age and multimorbidity, the number of patients who continue to require invasive ventilation after treatment of the acute condition rises without specialised weaning treatment. This leads to impaired quality of life, high costs and, above all, ties up human resources.

Objectives: The overall aim of this study is to reduce the number of patients who are classified as unsuccessfully weaned from invasive ventilation according to category 3cI of Germany's current guideline on prolonged weaning. Thus, the pri-mary endpoint is defined as successful weaning from invasive ventilation in high-risk patients. For this purpose, a prog-nosis model will be developed which will predict the risk of long-term invasive ventilation in patients who have already been ventilated for 96 hours along with a structured care concept (PRiVENT-intervention) that will be conceptualised during the preparatory phase of the study. Besides, predefined health economic aspects and the study's process evalu-ation will be described in a separate ethics proposal.

Methods: This is an intervention study with a parallel control group conducted in Baden-Württemberg over 24 months. Four weaning centres will each recruit 10 clinics or intensive care units, in total 40 clinics. Inclusion criteria for patients are: having required invasive ventilation for ≥96 hours, no more than seven days have passed since com-pletion of the 96th ventilator hour, being ≥30 years of age, suffering from at least one comorbidity or acute Covid-19 pneumonia while excluding neuromuscular diseases without weaning potential. The patients' risk of prolonged ventila-tion is assessed by the prognosis model. Independently, patients with acute Covid-19 pneumonia will be declared high-risk patients and treated in the intervention group. It is expected that about 1500 patients with an increased risk of long-term invasive ventilation will receive the complex intervention within the intervention period. In the control group, the expected number of high-risk patients is estimated to be the same. To compare the intervention with the current standard of care, AOK Baden-Württemberg's healthcare claims data will be assessed. Thereby, differentiating between healthcare claims data from all non-participating centres as well as participating centres in Baden-Württemberg.

Data analysis: The collected study data as well as AOK Baden-Württemberg's healthcare claims data will be analysed descriptively. Further, a mixed logistic regression model will be used to evaluate the primary endpoint, successful wean-ing from invasive ventilation. This model will incorporate random effects to control for the clustering effect of centres, as well as fixed effects adjusting for group membership (intervention vs. control), age, gender, Covid-19, previous diseases, number of patient cases requiring invasive mechanical ventilation at the treatment centre, and other variables identified by the prognosis model. Where appropriate, secondary endpoints will be explored in a similar way to the primary evalu-ation model.

Discussion: If the results of the study indicate improvement in the ventilation situation of patients otherwise requiring long-term invasive ventilation, implementation of the intervention into regular care will be sought.

Inclusion Criteria:

• ≥ 96 hours of invasive ventilation,
• no more than 7 days have passed since the patient completed their 96th hour of invasive venti-lation,
• ≥ 30 years old,
• ≥ 1 comorbidity and/or acute Covid-19 pneumonia
• not suffering from any neuromuscular disease without weaning potential.

Exclusion Criteria:

All patients who do not meet the above inclusion criteria and who do not give their consent to take part in the study will be excluded from participating in the study. The latter does not apply if a patient's legal representative will give consent on behalf of the patient as outlined in the study's inclusion criteria.