Understanding the diagnostic timeliness of cancer patients with pre-existing morbidities: What do different methodological approaches tell us?

Conference:

SAPC ASM 2023 - Brighton

Talk Code:

4A.1

Presenter:

Gary Abel

Twitter:

@garyabel

Co-authors:

Bianca Wiering, Sarah Price, David Shotter, Jose M. Valderas, Sam Merriel, Sarah Moore, William Hamilton, Luke Mounce

Author institutions:

University of Exeter, National University of Singapore, University of Manchester

Problem

Studies have suggested that cancer patients with pre-existing co-morbidities experience longer times between presentation in primary care and diagnosis (diagnostic interval) than patients without co-morbidities, potentially contributing to worse outcomes in these patients. However, establishing these timelines depends on the identification of an index consultation, which has the potential to introduce bias when comparing groups with different background consulting patterns. Here we aim to compare findings from traditional approaches of investigating diagnostic timeliness with an alternative approach based on trends in the rate of consultations prior to diagnosis.

Approach

Using linked primary care and cancer registration data for patients diagnosed cancer (2012-2018) we constructed 4 groups with varying multimorbidty burden using the Cambridge Multimorbidity Score. The diagnostic interval was calculated for all patients with a feature of possible cancer in the year before diagnosis. We also used a novel maximum likelihood based methods to estimate the time before diagnosis when population consultation rates increased (the inflection point) stratified by multimorbidity burden group.

Findings

The median diagnostic interval was 63 days which varied by multimorbidity burden; 35 days in those without pre-existing comorbidities; 135 days in the highest burden group. Contrastingly, the consultation rate inflection point varied little by morbidity burden; 126 days in those without pre-existing comorbidities or low/medium morbidity burdens; 112 days in those with high burden (p=0.054). Results by cancer site will also be discussed.

Consequences

Our findings that cancer patients with multimorbidity have longer diagnostic intervals concur with previous work using similar methodology. However, using a different approach we reach the conclusion that multimorbidity has little impact on diagnostic timeliness. We posit this difference can be explained by an artefactual bias in traditional approaches. Given these findings, explanations for worse outcomes in patients with pre-existing morbidities, other than delayed diagnosis, should be sought.

Submitted by:

Gary Abel

Funding acknowledgement:

This study/project is funded/part funded by the NIHR Programme Grants for Applied Research (PGfAR) SPOtting Cancer among Comorbidities (SPOCC) programme: supporting clinical decision making in patients with symptoms of cancer and pre-existing conditions (NIHR201070). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.