Risk of Cancer following Non-Specific Symptoms in Primary Care: Protocol for a Programme of EHR-Based Research in Primary Care
Open Access
- 23 October 2026
- journal article
- Published by F1000 Research Ltd in HRB Open Research
- Vol. 9, 107
Abstract
Background Non-specific symptoms of cancer (NSSoC), such as thrombocytosis, unexplained weight loss, and persistent fatigue, are frequently the earliest presenting feature of malignancy in primary care. Approximately 22% of cancers present with NSSoC alone, yet these symptoms have low individual positive predictive values and do not meet criteria for existing site-specific referral pathways. International evidence demonstrates that structured NSSoC pathways detect cancer in 7–20% of referrals. Ireland currently lacks both a national NSSoC referral pathway and the population-level epidemiological data required to inform one. Aim To conduct a programme of linked observational studies examining the prevalence, cancer conversion rates, and diagnostic intervals associated with NSSoC in Irish general practice. The initial pre-specified analysis focuses on thrombocytosis, using structured laboratory data to establish the analytical framework before extension to symptom-based NSSoC. Methods This protocol describes a retrospective cohort study using the CRADLE database (66 general practices; ~521,000 patients; 2.16 million person-years). Four integrated components will: (i) estimate NSSoC consultation-level burden, person-level annual incidence, and test positivity rates; (ii) characterise pre-diagnostic symptom trajectories via matched case-control diagnostic window analyses; (iii) estimate 6- and 12-month absolute cancer risk and cumulative incidence, accounting for competing mortality; and (iv) model adjusted cancer risk. Cancer outcomes are ascertained from GP-coded diagnoses; the absence of National Cancer Registry linkage is a primary limitation, and conversion rate estimates should be regarded as provisional. Implications This study will provide the first Irish primary care estimates of NSSoC burden and short-term cancer risk. Findings will generate candidate risk estimates to inform microsimulation modelling of a potential NSSoC pathway, contributing to the goals of the NCCP Early Diagnosis Plan 2022–2027. Definitive referral thresholds will require validation through cancer registry linkage and prospective evaluation.Funding Information
- Breakthrough Cancer Research (Innovation_001_2024)
This publication has 20 references indexed in Scilit:
- Patient characteristics, serious disease diagnoses, and incidental findings in individuals with non-specific symptoms referred to the Suspected CANcer (SCAN) Pathway: a prospective cohort study in EnglandThe Lancet Primary Care, 2025
- Artificial intelligence-aided data mining of medical records for cancer detection and screeningThe Lancet Oncology, 2024
- Cross-sectional study using primary care and cancer registration data to investigate patients with cancer presenting with non-specific symptomsBMJ Open, 2020
- Natural language processing of symptoms documented in free-text narratives of electronic health records: a systematic reviewJournal of the American Medical Informatics Association, 2019
- GRIPP2 reporting checklists: tools to improve reporting of patient and public involvement in researchBMJ, 2017
- Determinants of general practitioner's cancer-related gut feelings—a prospective cohort studyBMJ Open, 2016
- The expanding role of primary care in cancer controlThe Lancet Oncology, 2015
- Is increased time to diagnosis and treatment in symptomatic cancer associated with poorer outcomes? Systematic reviewBritish Journal of Cancer, 2015
- Rethinking diagnostic delay in cancer: how difficult is the diagnosis?BMJ, 2014
- Evidence of increasing mortality with longer diagnostic intervals for five common cancers: A cohort study in primary careEuropean Journal of Cancer, 2013