Figure 1

Flow diagram of patient selection and inclusion. Follow up was lost in 40 patients, usually because the patients did not live in the area of the hospital (they were on vacation). Characteristics of those patients did not differ from those of the group analyzed in the study (data not shown). A large group of patients (n = 1,229) were admitted to the intensive care unit (ICU) for under 48 hours and hence were excluded from the final analysis. Patients who died within 48 hours of ICU admission (n = 44) were excluded. In some cases the patient had no close proxy (n = 36). Patients re-admitted to the ICU were excluded (n = 132) because it was possible that the first admission could have biased the proxy memories of the patient's pre-admission health-related quality of life (HRQOL). Proxies or the patients themselves refused informed consent (n = 98) mainly because they felt study participation to be too great a burden at that stressful moment. Patients transferred to other hospitals (n = 16) or with cognitive impairment (n = 60), or who did not speak sufficient Dutch (n = 12) were also excluded. Some patients were not included because of investigator absence (n = 49). LOS, length of stay.