Prognostication of Need for Rehabilitation and Special Support in ICU Survivors

Every year, an estimated 5 million patients are treated in European Intensive Care Units (ICUs) yearly, with 40 000 in Swedish ICUs alone (1,2). As many as 50 % of ICU survivors suffer from new-onset physical, psychological or cognitive problems in the months or even years after ICU discharge, irrespective of the reason for ICU admission (3-5). These problems are major hurdles for patients discharged from the ICU to return to life as it was prior to hospitalisation, but often remain unrecognised by healthcare providers (3-5). Follow-up led by ICU clinicians is increasingly recognised as an addendum to critical care (5,6). Yet, interventions in large and unselected groups of ICU survivors have failed to demonstrate meaningful improvements (7,8) except in vulnerable subgroups (9,10). Dilution of treatment effects is a recognised problem in interventional studies (11). Thus, identifying patients with risk for adverse outcome is an important enrichment step prior to evaluation of new interventions. The importance has recently been highlighted at a stakeholders meeting regarding post-ICU syndrome (12). Pilot studies from our research group indicate that risk prediction instruments for ICU survivors with fairly good accuracy can be developed and used in clinical practice (13,14) but predictors need to be assessed and validated in a larger sample.

Data collection All data will be entered in coded form to a web-based, electronic Case Report Form (Pheedit eCRF) supplied by the Central Steering Committee. The eCRF will be accessible with a username and password at all computers with internet access. All data can be entered in the eCRF. Each centre will keep a safely placed and locked patient ID and study code list, separate from the study data.

Potential risk factors:

Patient characteristics Age Sex Educational level** Working/school/parental leave/retired/unemployed/sick-leave** Physical comorbidities (Charlson Comorbidity Index & SAPS III box 1) (supplement) History of psychological problems before ICU stay, including substance abuse) Caretaker of a child <18 years old** Barthel Index-100 prior to falling ill, reported by patient, and/or proxy together with medical chart review

In-ICU diagnoses/treatments/status Admission diagnosis (medical/surgical/trauma) Acute/elective hospital admission Admission severity of illness (APACHE II) Pelvic fracture Fractures (other than pelvic) ICU length of stay (hours) Severe sepsis/septic chock* Days with coma (for any reason) Days with invasive ventilator treatment Agitation/agitated delirium during ICU stay**

Patient status at ICU discharge Depressive symptoms with PHQ2 questions (two questions, supplement)** Chelsea Critical Care Physical Assessment Tool (CPAx) Questions 1-5 (supplement) within 24 hours prior to discharge* PTSS-14 Part A (four questions, supplement)** Social support (one question, supplement)

Three months post-ICU discharge Hospital Anxiety and Depression Scale (HADS) (supplement) Posttraumatic Stress Symptoms 14 (PTSS-14) (supplement) Barthel Index (supplement) Short Form 36 Health Survey (supplement) Health utilization questionnaire

Questionnaires will be sent by postal mail to all patients, together with an information letter about participation in the studies and a consent form to sign for those unable to give consent at inclusion. A phone number will be supplied, for patients with questions or who wish to know more regarding their ICU stay. Ten-fourteen days after questionnaires have been sent out, a reminder telephone call will be made to patients who have not returned the questionnaires.

Consent Patients will be asked for consent at study inclusion. If unable to consent, patients will be asked for consent at three months post ICU, when questionnaires are being sent out.

Inter-rater reliability Risk factors that are assessed by observation, where no validated instruments are available, will undergo interrater reliability testing within the study. Ten patients per centre will be assessed in parallel by two independent observers, blinded to each other´s assessment. Specifically, the factors are the items of CPAx.

Outcome measures:

Two primary outcome measures:

• Adverse psychological outcome at three months post-ICU discharge is defined as a subscale score >10 in the Hospital Anxiety and Depression Scale (HADS), indicative of clinically significant depressive or anxiety symptoms, AND/OR a Posttraumatic Stress Symptoms -14 (PTSS-14) part B score >45 indicative of clinically significant posttraumatic stress symptoms.
• Adverse physical outcome (new-onset physical disability) three months post-ICU discharge is defined as a worsening of the Barthel Index-100 by 10 points or more, compared to the reported Barthel Index-100 two weeks prior to hospitalisation.

Inclusion criteria

ICU stay>24 hours for postoperative patients and >12 hours for other ICU admissions

Exclusion criteria

• Previously documented cognitive impairment (e.g. dementia or intellectual development disorder resulting in poor understanding/compliance in outcome assessment)
• No home address
• Unable to read and write in language spoken at local study site (Swedish/Danish/Dutch)/used in questionnaires
• Patient declines participation
• Moribund patient or with more than one limitation of therapy
• Need for neurointensive care due to head trauma, intracranial hemorrhage or infarction AND GCS never better than 13 in the first 48 hours
• Transfer to other ICU before discharge to ward
• Solely in ICU for elective procedure (eg central line, epidural)