Changing healthcare utilisation patterns in cancer patients prior to their diagnosis: A literature review to help determine the length of potential diagnostic windows

Introduction

Promptly suspecting the diagnosis of cancer in patients presenting with new symptoms of cancer remains a challenge, and many patients are still diagnosed at a late stage. Ongoing research is needed to identify missed opportunities for earlier diagnosis, to target improvement efforts.

Studies using electronic health records have shown that in cohorts of patients subsequently diagnosed with cancer, consultation rates, and the use of diagnostic tests or prescriptions start to increase from baseline long before their diagnosis. It has been suggested that the onset of changes in healthcare utilisation rates defines the start of a ‘diagnostic window’, during which earlier diagnosis would in principle be possible. This highlights opportunities to diagnose at least some of the patients earlier, by better appreciating and acting on the ‘signals’ indicated by changing patient healthcare utilisation, or other signs and symptoms within the diagnostic window.

Currently there is no systematic appreciation of how much earlier cancer patients could be diagnosed in principle, as signalled by the onset of increasing healthcare use pre-diagnosis, and for which patients’ potential opportunities are greatest. Furthermore, the range of different types of healthcare utilisation events that could be used to identify the onset of diagnostic windows is unclear.

Funding

Cancer Research UK

Aims & objectives

To summarise evidence from population-based studies using electronic health records about the maximum length of reported ‘diagnostic windows’ indicated by changing healthcare utilisation before diagnosis, and examine related variation by patient group where applicable.

Methodology

A literature review of population-based studies based on electronic health records reporting pre-diagnostic changes in healthcare utilisation among patients subsequently diagnosed with cancer.

The range of ‘diagnostic windows’ across the studies by clinical event type will be summarised, for all cancers combined and by specific cancer site.

Related projects

Examining methodology to identify patterns of consulting in primary care for different groups of patients before a diagnosis of cancer: a matched retrospective cohort study

Publications

• White B, Renzi C, Rafiq M, Abel GA, Jensen H, Lyratzopoulos G. Does changing healthcare use signal opportunities for earlier detection of cancer? A review of studies using information from electronic patient records, Cancer Epidemiology, 2021, 102072, ISSN 1877-7821, doi: 10.1016/j.canep.2021.102072
• White B, Rafiq M, Gonzalez-Izquierdo A, Hamilton W, Price S, Lyratzopoulos G. Risk of cancer following primary care presentation with fatigue: A population-based cohort study of a quarter of a million patients. Br J Cancer (2022). doi: 10.1038/s41416-022-01733- 6