Management of upper gastrointestinal hemorrhage occurring in the general population: patterns of health care management in four French geographical areas

Aims - To describe patterns of health care management in patients with uppe r gastrointestinal hemorrhage and to identify factors linked to the differe nt patterns. Patients and methods-We conducted a prospective study of patients over 18 w ith upper gastrointestinal hemorrhage (inpatients excluded) among all publi c hospitals and private practice gastroenterologists in 4 French administra tive areas (3 in Northern France and one in the South West). Results - One thousand six hundred and two patients were included over a si x-month period (1996). An endoscopic procedure was performed in 1532 patien ts in public (70%) or private (20.5%) hospitals, or at private office (9.5% ). Hospitalization was necessary in 78.8% of the patients in universitary, non universitary public or private hospitals (38.9, 45.5 and 15.6%, respect ively) with a median duration of 6.5 days. Admission was associated to old age, short delay between hemorrhage and endoscopic procedure previous gastr ointestinal bleeding, cirrhosis or cancer bleeding from peptic ulcer or eso gastric varices. Endoscopic hemostasis was performed in 21.4% of the patien ts, more often in universitary and no universitary public hospitals. Surger y was necessary in 4% of the patients. Death rate was 10.7%. Important geog raphical variations were observed concerning referral patterns. Patients' c haracteristics did not differ between the 4 areas. On the other hand, healt h care supply provided in the management of upper gastrointestinal hemorrha ge was different in the four French geographical areas. Conclusion - a) An initial endoscopic procedure is nearly always performed in patients with an upper gastrointestinal hemorrhage in France; in 1 patie nt out of 10, endoscopy was performed in a private gastroenterologist offic e; b) hospital admission was strongly related to epidemiological and clinic al criteria of severity; c) the geographical variations observed in referra l patterns depend in part on health care supply; d) upper gastrointestinal haemorrhage status could be used as an indicator of the quality of health c are organizations.