Clinician estimates of survival in palliative care patients

• Main Author: White, Nicola Gayle
• Other Authors: Stone, P. ; Harris, A. ; Harries, P.
• Published: University College London (University of London) 2017
• Subjects: 616.89
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.746887

Doctors need to identify when palliative care patients are imminently dying so they can adjust their goals of care and treatment accordingly. The systematic review of the literature, completed as part of this thesis, showed that these decisions are very inaccurate. The aim of this thesis was to determine the judgement policies of expert prognosticators (i.e. those clinicians with a proven ability to identify patients in the last 72 hours of life). This information is needed to help the development of training programmes for less expert clinicians. In order to understand decision-making judgements, it was first necessary to identify an “expert” group of clinicians. Previous studies suggested that expertise is not defined by years of experience or seniority. Therefore to identify an expert group with proven prognostic abilities, palliative care doctors (n = 99) completed a prognostic “test”. The test was developed by compiling case histories from the direct observation of 50 seriously ill patients in a hospice and an acute hospital. In order to complete the test, doctors were asked to review 20 case summaries and to provide a percentage likelihood that each patient would die within the next three days. The top 20% of doctors who performed most accurately on this test were deemed to be “expert prognosticators” and were invited to participate in the next phase of the research. The expert group (n = 19) were asked to complete a further prognostic task so that the decision-making policy of each individual (and the expert group as a whole) could be determined using Judgement Analysis. Through statistical modelling, experts had weighted Cheyne-Stoke breathing (β=15.44), the standardised palliative performance score (β=12.35) and the rapidity of decline in the previous 24 hours (β=11.512) as the most important factors. They gave lower weighting to the standardised level of agitation and sedation (β=5.97), the presence of audible secretions (β=5.95) and the presence of cyanosis (β=5.38).