TY - JOUR T1 - OP28 Association of diagnostic intervals with breast, prostate, lung and colorectal cancer survival in England: historical cohort study using the Clinical Practice Research Datalink JF - Journal of Epidemiology and Community Health JO - J Epidemiol Community Health SP - A16 LP - A17 DO - 10.1136/jech-2014-204726.31 VL - 68 IS - Suppl 1 AU - MT Redaniel AU - RM Martin AU - M Ridd AU - J Wade AU - M Jeffreys Y1 - 2014/09/01 UR - http://jech.bmj.com/content/68/Suppl_1/A16.2.abstract N2 - Background Rapid diagnostic pathways were implemented within the UK National Health Service (NHS) with the aim of improving cancer outcomes and increasing cancer survival in the UK. The evidence whether shorter diagnostic intervals (time from primary care presentation to diagnosis) do translate to improved survival, and whether this differs by presenting symptom (s), is unclear. Methods Using the Clinical Practice Research Datalink (CPRD), we identified patients diagnosed with breast (female, 8639), colorectal (5928), lung (5846) and prostate (2807) cancers between January 1, 1998 and December 31, 2009, and aged 15 years and older at the time of diagnosis. We used relative survival and excess risk modelling to determine associations between diagnostic intervals and five-year survival. All analyses were stratified by the presenting symptom: highly suggestive of cancer (‘alert’ symptoms, according to NICE guidance), or less directly suggestive but predictive of cancer (non-alert symptoms). Results Five-year relative survival was 80.5% for breast (95% Confidence Interval, 95% CI: 79.2–81.7%), 78.3% for prostate (95% CI: 75.5–81.0%), 48% for colorectal (95% CI: 46.3–49.7%) and 7.9% for lung cancer (95% CI: 7.1–8.8%). The survival of patients with colorectal, lung and prostate cancer was greater in those who presented with alert, compared with non-alert, symptoms. For patients with breast cancer, survival was greater in those presenting with non-alert symptoms. Longer diagnostic intervals were not associated with higher excess mortality amongst breast and lung cancer patients with alert symptoms. There was some evidence that both short (<1 month) and long (>6 months) diagnostic intervals were associated with increased mortality amongst colorectal patients with alert symptoms, compared with intervals of 1–2 or 3–6 months. For colorectal and lung cancer patients with non-alert symptoms, longer diagnostic intervals were associated with lower mortality (colorectal cancer: Excess Hazards Ratio, EHR >6 months versus <1 month: 0.85; 95% CI: 0.72–1.00; Lung cancer: EHR 3–6 months versus <1 month: 0.87; 95% CI: 0.80–0.95; EHR >6 months versus <1 month: 0.81; 95% CI: 0.75–0.89). Prostate cancer survival increased with longer diagnostic intervals, regardless of type of presenting symptom. Conclusion The relationship between diagnostic intervals and cancer survival is more complex than had previously been assumed, varying by symptom site and cancer: in colorectal, lung and prostate cancer, patients with non-alert symptoms had poorer survival compared to those with alert symptoms, but their mortality rates were paradoxically lower as diagnostic intervals increased. The nature of presenting symptoms should be taken into account when studying the association of diagnostic intervals with survival. ER -