Prediction of psychological and physical morbidity after critical illness and intensive care unit stay

Sammanfattning: A large proportion of survivors of critical illness and intensive care unit stay suffer from post- intensive care syndrome (PICS), consisting of psychological, physical and cognitive problems. These problems can persist for months to years and impede the return to life as it was prior to falling ill. Psychological problems such as depression, anxiety and post- traumatic stress (PTS), can also affect informal caregivers to ICU patients. In order to detect and hopefully treat these problems, an increasing number of hospitals are offering follow-up in the months after ICU stay. Resources for ICU follow-up are limited and it is unknown which patients are at the highest risk of developing PICS. ICU length of stay longer than 3-4 days is currently the most commonly used and recommended method for selection of patients for follow-up, but evidence is lacking regarding the accuracy of this method in finding patients with the highest risk for PICS. The aim with this thesis was to assess risk factors for psychological and physical sequelae after ICU stay, and to develop instruments to predict individual patients’ risk of these adverse outcomes, as well as investigate the effect of patient outcome on the wellbeing of their informal caregivers. Study I is a multicenter prospective observational cohort study assessing risk factors for psychological problems (PTS, depression and anxiety) three months post-ICU in order to develop a discharge screening instrument for identification of patients for psychological ICU follow-up. We included 572 patients at ten ICUs in Sweden, Denmark and the Netherlands. Among 404 (78%) responders, 20% developed significant symptoms of any of the assessed psychological entities. After univariable and multivariable logistic regression modeling, the remaining predictors for adverse psychological outcome three months post-ICU were: age (with the highest risk in ages 49-65 years), lack of social support, symptoms of depression and traumatic memories at ICU discharge. The area under the receiver operating characteristics curve (AUC) for the screening instrument was 0.76 (95% CI 0.70-0.81). Study II is a multicenter prospective observational cohort study assessing risk factors for new-onset physical disability three months post-ICU in order to develop a discharge screening instrument for identification of patients for physical ICU follow-up. Included patients are the same as in study I. Among the 404 responding ICU survivors, 19% reported new-onset physical disability. After univariable and multivariable logistic regression modeling, the sole remaining predictor for an adverse outcome was physical status at ICU discharge, with an AUC of 0.68 (95% CI 0.61-0.76). Study III is a single-center prospective observational cohort study evaluating the predictive value of an early psychological assessment one week after ICU discharge on three-month psychological outcome regarding symptoms of PTS, depression and anxiety. Among 132 included patients, there are follow-up data on 82 (62%). In our cohort, 13% suffered from clinically significant symptoms of PTS, 21% from symptoms of depression and 16% from symptoms of anxiety at three months. Correlation between early scores in the ward and three months scores were moderate to strong. The predictive value of the early screening as assessed with the AUC was 0.90 (95% CI 0.81 to 0.99) for symptoms of PTS, 0.80 (95% CI 0.64 to 0.95) for symptoms of anxiety and 0.75 (95% CI 0.64 to 0.87) for depressive symptoms. Study IV is a multicenter prospective observational cohort study including cohabiting informal caregivers to 62 ICU survivors included in study I/II. The primary outcome was to assess whether an adverse psychological or physical patient outcome was associated with a higher degree of caregiver burden three months post-ICU. Response rate was 89% (n=55). Of included patients, 17 (33%) had an adverse outcome. Caregiver burden was significantly higher in caregivers to patients with an adverse outcome, caregiver burden scale score mean (±SD) 52 (11) compared to caregivers caring for patients without an adverse outcome, mean 41 (13), p=0.006. A higher caregiver burden also correlated with a reduced caregiver mental health-related quality of life. This thesis developed methods for prediction of psychological and physical sequelae in ICU survivors three months post-ICU, as well as assessed the effect of an incomplete patient recovery on caregiver burden and mental health problems in informal caregivers. The results of this thesis provides clinicians with tools to better anticipate the trajectory of recovery for their patients in order to initiate early interventions in high-risk patients and their informal caregivers and potentially reduce long-term suffering. This triage of patients also allows for enrichment of high-risk cohorts for future interventional studies of ICU follow-up interventions.