Article Text


BSG endoscopy section free papers
OC-044 Gastroscopy rate in English general practice populations: association with outcome for oesophagogastric cancer
  1. M Shawihdi1,
  2. G Powell1,
  3. N Stern2,
  4. N Kapoor2,
  5. R Sturgess2,
  6. E Thompson1,
  7. M Pearson1,
  8. K Bodger3
  1. 1Aintree Health Outcomes Partnership, University of Liverpool, Liverpool, UK
  2. 2Digestive Diseases Centre, Aintree University Hospital, Liverpool, UK
  3. 3Department of Gastroenterology, Institute of Translational Medicine, University of Liverpool, Liverpool, UK


Introduction Rates of gastroscopy vary between English general practice populations. The magnitude of this variation suggests a wide spectrum of clinical practice. Current guidelines focus on alarm symptoms as triggers for investigation but early symptoms of cancer are non-specific. This project aimed to determine whether overall gastroscopy rate in GP practice populations in England is associated with outcome of oesophagogastric cancer (OGC), as measured by rate of major surgical resection, emergency admission for cancer diagnosis and mortality.

Methods Analysis of Hospital Episode Statistics (HES, 2006–2008) linked to death registry and practice population data. Gastroscopy volume determined by extracting total diagnostic gastroscopy procedures and aggregated at GP practice level. OGC cases: Methods developed and validated (using local & national audit) to identify new cases of OGC and then extract all hospital episodes in chronological order, flag key milestones (eg, diagnostic gastroscopy; emergency admission to hospital; major surgery) using relevant diagnostic and procedure codes. Entry criteria: General practices with ≥1 new case of OGC and with a per capita gastroscopy rate within a valid reference range (0.4–4.0 per 1000 population). Practices grouped into tertiles (low, medium and high gastroscopy rate).

Results 20 709 OGC cases from 5956 practices serving an adult population of 35.1 million. Characteristics of OGC cases matched the national audit findings. Cases registered with practices in lowest tertile of gastroscopy rate had lowest rate of surgery (14% vs 16% vs 16%; p=0.028), highest rate of emergency admission (29% vs 27% vs 25%; p<0.01), and highest mortality at 6 months (41% vs 40% vs 39%; p<0.01). After adjustment for age, sex, co-morbidity and deprivation quintile in logistic regression analysis, the rate of gastroscopy (low, medium or high) at the patient's general practice was an independent predictor of all three outcomes.

Conclusion There is >10-fold variation in the rate of gastroscopy among general practice populations in England. On average, OGC patients belonging to practices within the lowest tertile have poorer outcomes. These findings suggest that guidelines aimed at reducing the use of gastroscopy may adversely affect cancer outcomes.

Competing interests None declared.

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.