Influence of the hospital upon the outcomes of gastrointestinal fistulas: results of a multi-continent, multi-national, multi-center cohort

RationalePrognosis and outcomes of gastrointestinal fistulas (GIF) might depend upon the operational characteristics of the hospital containing and caring for the patients. ObjectiveTo assess how selected operational characteristics of the hospital participating in the exercises of the "Fistula Day" Project (FDP) influence upon prognosis and outcomes of GIF. Study designCohort-type study. Enrolled patients were followed for 60 days. Three cross-sectional examinations were made during the completion of the exercises of the FDP, namely, upon admission of the patient in the study, and 30 and 60 days after admission. Study serieSeventy-six hospitals of Latin America (13 countries) and Europe (4). MethodsAssociations between survival of the patient, prolongation of hospital stay, and (likely) spontaneous closure of the fistula, on one hand; and selected operational characteristics of the participating hospital, on the other; were assessed. ResultsSpecialties hospitals prevailed. Most of the hospitals assisted between 1 - 2 GIF patients a month. Participating hospitals distributed evenly regarding the number of beds. Most of the hospitals had an intensive care unit. Similarly, three-quarters of the hospitals had a multidisciplinary unit dedicated to clinical and hospital nutrition. However, a unit dedicated to the management of intestinal failure and/or postoperative fistulas was present only in a fifth of them. Experience of the physician attending GIF was rated between "Expert" and "High" in one third of the hospitals. Number of hospitals beds associated with increased survival of GIF patients ({chi}2 = 5.997; p = 0.092), prolonged hospital stay ({chi}2 = 7.885; p < 0.05), and higher rate of spontaneous closure of the fistula ({chi}2 = 11.947; p < 0.05). In addition, rate of spontaneous closure of the fistula was (marginally) higher among patients assisted by a hospital unit specialized on intestinal failure ({chi}2 = 3.610; p = 0.0574). On the other hand, survival of the patient was dependent (also marginally) upon the number of patients assisted in a month ({chi}2 = 5.934; p = 0.0514). ConclusionsIt is likely number of hospital beds to determine prognosis and outcomes of GIF. Other operational characteristics of the hospital might exert a marginal influence upon survival of the patient and the likely spontaneous closure of the fistula.