Abstract

Continuing reductions in cardiovascular disease mortality have brought about a new type of epidemiological transition, characterised by increasing disease burden associated with cancer. In spite of many critical breakthroughs in cancer treatment, prognosis for advanced stage cancer remains poor for most cancers. Screening interventions with proven efficacy and cost-effectiveness remain stubbornly limited to few cancers (such as bowel and cervical cancer) whilst debate about “over-treatment vs mortality benefit” trade-offs bedevil breast and lung cancer screening. These realisations have sharply focused attention to efforts to diagnose cancer as soon as symptoms present, in the hope of diagnosing cancer at an earlier stage. Nonetheless, cancer symptoms (as at least understood currently) have low to very low specificity and predictive value. In this lecture I will summarise a range of epidemiological studies on variation in diagnostic intervals before and after presentation, together with evidence on variation in markers of diagnostic timeliness (such as diagnosis through an emergency presentation or stage at diagnosis). These findings provide insights into the complex psychological factors associated with late help-seeking for symptoms likely to be due to cancer; and mechanisms likely to be responsible for delays in suspecting cancer after presentation. I will conclude with recommendations for better targeting of policy initiatives and research that could help to bring about more timely diagnosis of cancer for the greatest possible number of current and future cohorts of cancer patients.