P80 Validating bristol prognostic score for predicting 1year mortality an objective tool for palliative care referral

<h3>Background and Aim</h3> Early referral to palliative care and subsequent advanced care planning have been highlighted as an area of unmet need in patients with advanced chronic liver disease (ACLD). Fluctuating disease trajectory and uncertain prognosis have been described as challenges to physicians initiating palliative care discussions. The Bristol Prognostic Score (BPS) was developed to help identify patients with a predicted 12-month mortality to guide referral to palliative care. Studies have shown regional variations in the performance of BPS. We aimed to assess its performance in our cohort of patients prior to embedding its use in our liver admission bundle <h3>Method</h3> Retrospective data was collected for consecutive hospital admissions for a period of 5 months (Jan 2019 to May 2019) admitted with an episode of decompensated ACLD. Electronic patient records (Meditech®) were interrogated to curate patient demographic data and relevant variables of interest. The primary outcome was validation of BPS in predicting all-cause 1-year mortality. <h3>Results</h3> 63 patients had 97 inpatient admissions in the 5-month period. The most common aetiology was alcohol related cirrhosis (85%) with mean age of 56 (SD±13) and a male preponderance 79% (50/63). 30 (48%) had a BPS ≥ 3 and one third of all patients died with 12 months of index admission episode. 3 patients (all with BPS &lt;3) went on to have a liver transplant and were therefore excluded from further analysis. A BPS threshold of ≥3 predicted 1-year mortality with a sensitivity of 57.1%, specificity of 53.8%, and positive predictive value (PPV) of 40.0%. However, when considering 3-year mortality the same threshold yielded improved performance (sensitivity 55.3%, specificity 59.1%, PPV 70.0%). <h3>Conclusion</h3> In our observation, BPS did not perform as well in predicting 1 year mortality compared to the original study (sensitivity 72.2%, specificity 83.3%, PPV 81.3%). Our results were similar to Scottish hospitals data (sensitivity 54.9%, specificity 58%, PPV 43.4%). However, it’s performance was improved when predicting 3-year mortality and may be useful in complementing clinical judgement when considering prognosis and palliative care referral in patients with ACLD. <h3>References</h3> Woodland H, Hudson B, Forbes K, On Behalf of the British Association for the Study of the Liver (BASL) End of Life Special Interest Group, <i>et al</i>. Palliative care in liver disease: what does good look like? <i>Frontline Gastroenterology</i> 2020;11:218–227. Hudson BE, Ameneshoa K, Gopfert A, <i>et al</i>. Integration of palliative and supportive care in the management of advanced liver disease: development and evaluation of a prognostic screening tool and supportive care intervention. <i>Frontline Gastroenterology</i> 2017;8:45–52. Bowers SP, Clare K, Hagerty L, <i>et al</i>. Predicting 1-year mortality among patients with decompensated cirrhosis: results of a multicentre evaluation of the bristol prognostic score. <i>BMJ Open Gastroenterology</i> 2022;9:e000822. doi: 10.1136/bmjgast-2021–000822