The frequency, nature and impact of GP-assessed avoidable delays in a population-based cohort of cancer patients

Background: There is a growing emphasis on the speed of diagnosis as an aspect of cancer prognosis. While epidemiological data in the last decade have quantified diagnostic timeliness and its variation, whether and how often prolonged diagnostic intervals can be considered avoidable is unknown.

Methods: We used data from the English National Cancer Diagnosis Audit (NCDA) on 17,042 patients diagnosed with cancer in 2014. Participating primary care physicians were asked to identify delays in diagnosis that they deemed avoidable, together with the ‘setting’ of the avoidable delay and key attributable factors. We used descriptive analysis and regression frameworks to assess validity and examine variation in the frequency and nature of avoidable delays.

Results: Among 14,259 patients, 24% were deemed to have had an avoidable delay to their diagnosis. Patients with a reported avoidable delay had a longer median diagnostic interval (92 days) than those without (30 days). Of all avoidable delays, 13% were deemed to have occurred pre-consultation, 49% within primary care, and 38% within secondary care. Avoidable delays were mostly attributed to the test request/performance phase (25%). Multimorbidity was associated with greater odds of avoidable delay (OR for 3+ vs no comorbidity: 1.43 (95% CI 1.25–1.63)), with heterogeneous associations with cancer site.

Conclusion: We have shown that GP-identified instances of avoidable delay have construct validity. Whilst the causes of avoidable diagnostic delays are multi-factorial and occur in different settings and phases of the diagnostic process, their analysis can guide improvement initiatives and enable the examination of any prognostic implications.