Abdominal pain (AP) is a symptom commonly reported to United Kingdom (UK) general practitioners (GPs). It can be a symptom of both transient illness and serious disease; often making its cause difficult for GPs to diagnose. This ambiguity influences GP decision making, such as where to refer the patient for further investigation - thus potentially causing delays in diagnosis. Existing research has shown abdominal pain to be a prior symptom of nine different cancers.
The aim of this research is to examine new cases of unexplained abdominal pain and to a) identify how many go on to have a cancer diagnosis in the next two years; b) categorise which cancers are diagnosed; c) investigate what co-symptoms are present for each cancer site and their impact on the chance of having cancer; d) use the results to recommend a best practice method of investigating abdominal pain; and, e) report what other diagnoses are found.
The results will help to guide GPs into providing appropriate action for patients with unexplained abdominal pain. They will also help to inform future policy on how best to investigate abdominal pain: in the GP surgery using ultrasound or in a specialist diagnostic clinic. New specialist clinics have been established, but have no relevant evidence to help them select the best sequence of tests for patients.
The revised 2015 National Institute for health and Care Excellence (NICE) guidelines highlight abdominal pain as a feature of several cancers reported in the 12 months before the diagnosis. This study aims to quantify the overall and individual cancer site risk of unexplained new abdominal pain episodes reported to primary care in the 24 months following the AP episode, stratified by age and gender, and by accompanying symptoms. There is a need for a more effective diagnostic pathway, which may include primary care testing or extend to investigation in the newly-established multidisciplinary diagnostic centres (MDC).
1) To identify any new diagnosis of cancer in the two years (0-12 months and 13-24 months) following the first consultation with abdominal pain. The primary focus will be in the twelve months following the AP record, as the presence of AP is more likely to be related to the cancer than later diagnoses. A diagnosis of cancer will be ascertained by searching the clinical and referral files for medcodes (mapped to Read codes within the CPRD) relating to cancer. There are 2,134 medcodes relating to cancer within the CPRD. The diagnosis will be confirmed using cancer registry linkage. Information on cancer stage and grade will also be extracted. A sub-analysis of early versus late stage cancers will show if any investigation regime captures cancers preferentially. The date of diagnosis will be taken as the first recorded cancer code; where disparity between the CPRD and cancer registry exist, registry data will be prioritised.
2) What clinical features predict cancer in the abdominal pain group? (test: LR of cancer vs no cancer)
3) In the cancer group: Which specific features predict certain cancers a) firstly single cancers and b) cancers grouped by the diagnostic test of choice?
4) Alternative diagnoses. In patients investigated for cancer, alternative explanations for AP may be found. We have assembled a short list of plausible conditions which are likely to be identified serendipitously by cancer investigation. This outcome is what other plausible causes of abdominal pain may be identified by cancer investigation in this cohort?
Elizabeth Shephard - Chief Investigator - University of Exeter
William Hamilton - Corresponding Applicant - University of Exeter
NCRAS Cancer Registration Data