Is Invasive ICU-treatment Associated With Mental Illness?

Background As intensive care unit (ICU)-mortality rates are improving in many areas, research focus has turned towards sequelae after ICU-care. The term Post Intensive Care Syndrome has been coined, concluding a high frequency of mental illness, cognitive and physical impairment immediately after ICU-care. Since patients in the ICU experience life threatening conditions, severe psychological and physical stress, we hypothesize that patients after ICU have an increased risk of long-term mental illnesses that increase with the extent of intensive care, specifically anxiety, and post-traumatic stress disorder (PTSD). Current data on the extent and severity of mental illness after intensive care, and if intensive care is an independent predictor of it, is limited. Objective Our objective is to assess in nationwide data if extensive ICU-treatment contributes to mental illness. Extensive ICU-treatment will be defined in our dataset as ICU-treatment of patients in need of invasive ventilation, and/or continuous renal replacement therapy (CRRT). Our endpoints are prevalence of mental illness one year after discharge, and its association between extent of ICU-treatment. Method We will use the Swedish intensive Care registry of intensive care patients treated between 2010 and 2015 for extracting data on Data on Invasive ventilation treatment, CRRT-treatment, gender and age. ICD-10 codes corresponding for anxiety, PTSD and depression, and other diagnoses known to increase the risk of mental illness will be extracted from the Swedish National Patient Registry. Patients younger than 18 years old and ICU-codes corresponding for pre-existing anxiety, PTSD, depression, and treatment periods less than 24 hours will be excluded from the study. Missing data will be imputed. A multivariable regression model will be fit using ICU-codes corresponding with mental illness as dependent variable, Invasive ventilation and CRRT as independent variables, and the other parameters as covariates.