About this Research Topic
In 2018, there were 17 million new cases of cancer worldwide. For most patients, entry into the health system occurred through primary care following symptomatic presentation at general practice. In most countries, general practitioners (GPs) act as ‘’gatekeepers” to secondary health services and specialist care, assessing risk and ensuring appropriate and timely movement of patients from primary to secondary care. Based on the presenting problem or patient needs, GPs can pursue a range of actions, from issuing prescriptions, undertaking further investigations such as pathology, radiology or imaging services and referral to specialist care services. General practice can therefore make important contributions to improving cancer outcomes through early recognition and investigation of patients most likely to have the disease.
One of the most significant reasons for delays to diagnose cancer in primary care is that a GP may simply not consider cancer in their differential diagnosis. A full-time GP will only see 5-10 new cases of non-cutaneous cancer amongst their several thousand consultations per year. This low prevalence of cancer in primary care means that even so-called ‘red flag’ cancer symptoms have low positive predictive values.
Research using routine general practice data has helped inform our understanding of the epidemiology of cancer symptoms in primary care, so we can now estimate the likelihood of cancer based on a patient’s combinations of symptoms, signs and baseline risk factors. This has used traditional epidemiological approaches to identify predictors of a cancer diagnosis. However, this approach may miss potentially important additional signals and alternative methods to analyzing complex electronic GP data might provide novel insights to identifying patients with undiagnosed cancer. Current risk prediction models have not examined in detail, for example, trends in results of pathology tests, nor patterns over time of symptoms and drug prescribing. There is also not much practice of using the patient’s family medical history to determine patients with a greater risk of cancer, as genetic testing and counselling are not common primary care practices.
The Research Topic is intended to assign focus on latest machine learning and data mining techniques as well as advanced rule-based, deterministic methods, specifically ones that investigate possible new predictors for a future cancer diagnosis, assess trends in the results of common pathology tests, separately and in combination with reasons for attendance and prescription of certain medications. These methods, along with a patient's demographics and personal/family medical history, will further investigate the sections in the patient’s primary care data that can be used for a more targeted high risk and early cancer detection . We welcome all article types exploring the use of primary care datasets for high risk and early cancer detection.
Topic Editor Dr. Andrey Kan is an applied scientist for Amazon Inc. All other Topic Editors have no conflicts of interest to declare.
Keywords: primary care datasets, cancer detection, early intervention, machine learning, data mining, health informatics
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.