Do critically ill liver patients experience negative bias? A web-based survey examining doctors opinions to critical care escalation.

ObjectiveTo test the hypothesis that there is negative bias towards escalating levels of care in decompensated cirrhosis, compared with other patient groups.DesignAn electronic survey containing eight acute clinical scenarios with equivalent physiological derangement, in which respondents were asked to score the degree to which they would advocate for intensive care unit admission on a scale of 1-10. Scenarios included respiratory, haematology, vascular, renal, gastrointestinal, postoperative and hepatological conditions. Follow-up questions examined the reasons why the patient should or should not be transferred, and enquired about ceilings of care, end-of-life decisions, degree of organ support and healthcare financial rationing. 273 doctors responded.SettingSecondary care hospitals in south of England.PatientsNone involved.InterventionsNone.Main outcome measuresAdvocacy score (1-10) and subsidiary responses.ResultsThe hepatology patient ranked 4th of 8 with a mean advocacy score of 7.2. There were no significant differences between intensivists and physicians or between grades of seniority. Of those less likely to escalate (score 1-5, n=42), the reasons given were based on unsurvivability or excessive burden of treatment rather than aetiology. One-fifth cited 'lifestyle decision'. 25 (62.5%) respondents not favouring escalation would make the patient DNACPR, 17 (42.5%) would stipulate ward-based care only and a small minority would instigate active palliation. Of those favouring escalation (advocacy score 6-10), 70% (n=122) would consider unlimited organ support. Fifty-four (29.5% of those who answered) said they 'sometimes' or 'frequently' consider resource allocation when making decisions about escalation of care.ConclusionsWhen compared with a variety of acute medical scenarios, doctors did not overly appear to exhibit therapeutic nihilism for patients with decompensated liver disease; however, significant variation in interpretation of the data and management approaches was identified.