What attendance-level demographic, clinical, and temporal factors are associated with unscheduled care use in people who die from cancer?

Talk Code: 
3E.4
Presenter: 
Sarah Mills
Twitter: 
Co-authors: 
Author institutions: 
University of Dundee

Problem

Access to care out-of-hours is essential for supporting patients with advance cancer in living, and dying, well in the community. In the UK, unscheduled care is delivered by GP-Out-Of-Hours(GPOOH) and Accident & Emergency(A&E). Little is known about what factors influence use of unscheduled care by cancer decedents, and most of the research that does exist does not examine GPOOH.

Approach

This retrospective cohort study, of all 2,443 people who died from cancer in Tayside, Scotland from 03/2013-06/2015, aimed to understand the patterns and outcomes of unscheduled care attendance in the last 12 months of life, at GPOOH and A&E, by people who die from cancer, and to examine the associations with demographic, clinical, temporal, and prescribing factors. Clinical, demographic, cancer registry and prescribing datasets were linked to routinely collected clinical data using the Community Health Index(CHI) number. Anonymised data were analysed in SafeHaven. Analysis used Generalised Estimating Equation (GEE).

Findings

The majority of unscheduled care attendances, by cancer decedents in their last year of life, took place in GPOOH, rather than A&E. Cancer decedents who were women, younger, and had GI symptoms or infections had greater odds of using GPOOH; those who were male, older, had lung cancer, and presented with pain, breathlessness, acute neurological symptoms, or ‘unwell’ or ‘palliative care’, had higher odds of using A&E compared. Presenting complaints were associated with age and cancer type. Cancer decedents who attended A&E had over eight times greater odds of being admitted to hospital than those who attended GPOOH. Frequent or very frequent attenders, had greater odds of using GPOOH preferentially to A&E. Older cancer decedents, those coded as ‘unwell’ or ‘palliative care’, and those who attended after their cancer diagnosis, had lower odds of being admitted to hospital after unscheduled care attendance. Cancer decedents with a late diagnosis, those presenting with breathlessness, GI symptoms and infection, and those whose attendances had high clinical priority had greater odds of being admitted following unscheduled care attendance. Cancer decedents who were frequent or very frequent attenders had lower odds of being admitted to hospital following unscheduled care contact.

Consequences

This research demonstrates that unscheduled care use by cancer decedents is more prevalent than previously thought, and mostly delivered in GPOOH. It identified a number of factors associated with increased unscheduled care use. Targeting modifiable factors associated with increased unscheduled care use can minimise potentially avoidable unscheduled care use, and its consequences, for people dying from cancer. Identifying individuals with non-modifiable factors, who are at high risk of unscheduled care use – particularly frequent unscheduled care use – allows for focusing of resources and support to those people for whom it would confer the most benefit.

Submitted by: 
Sarah Mills
Funding acknowledgement: 
Sarah Mills was funded by a Clinical Academic Fellowship through the Chief Scientist Office (reference number: CAF_17_06). Initial funding for data extraction and storage was provided through PATCH Scotland Research Grants.