Background Patients with a diagnosis of chronic obstructive pulmonary disease (COPD) have a limited life expectancy and poorer quality of life, which can manifest with a high symptom burden especially towards the end of life. The benefits of offering palliative care to this patient group is well known but they frequently have poorer outcomes and worse access to palliative care services and symptom management. Knowing when to instigate appropriate palliative care in this group is challenging due to prognostic uncertainty. The Gold Standard Framework (GSF) is recommended for use in primary care settings with this patient group to predict 12 month mortality, it is stated to be accurate in 70% of cases, however there is little evidence to support this. This study aims to evaluate the effectiveness of this tool in predicting 12 month mortality in patients with a diagnosis of COPD.

Method Retrospective data from one community respiratory service in the North West of England, was collected between January 2013 and January 2016, of patient’s deemed to be entering the last 12 months of life by triggering the GSF tool. The length of time on the palliative register until either death or January 2107, when data analysis commenced, was reviewed. All the data utilised in this evaluation was routinely collected by the service.

Results Of the 448 patients included in this evaluation only 122 died within 12 months, meaning the GSF was accurate in predicting 12 month mortality in 27% of cases included in this review (n=448). Sub group analysis of the different GSF triggers and routinely collected data hints at the possible importance in pulmonary rehabilitation to reduce mortality (OR 0.1079 (95%CI 0.0674–0.1726 p<0.0001)) with gait speed potentially being an early prognostic indicator (p-value 0.008).

Conclusion This study suggests that in this cohort the GSF PIG is not an accurate predictor of 12 month mortality in patients with a diagnosis of COPD to the suggested 70%. It does hint at the importance of PR and gait speed in prognostication.