Jonathan Giovannelli1, Vincent Coevoet2, Chloé Vasseur3, Audrey Gheysens4, Blandine Basse5, François Houyengah6. 1. Univ Lille Nord de France; Epidemiology Unit, University Hospital of Lille, EA 2694 Lille, France. Electronic address: jonathan.giovannelli@gmail.com. 2. Medical Information Department, Dunkerque Hospital, Dunkerque, France. Electronic address: vincent.coevoet@ch-dunkerque.fr. 3. Nutrition Department, Dunkerque Hospital, Dunkerque, France. Electronic address: service.dietetique@ch-dunkerque.fr. 4. Medical Information Department, Dunkerque Hospital, Dunkerque, France. Electronic address: audrey.gheysens@ch-dunkerque.fr. 5. Nutrition Department, Dunkerque Hospital, Dunkerque, France. Electronic address: Blandine.Basse@ch-dunkerque.fr. 6. Medical Information Department, Dunkerque Hospital, Dunkerque, France. Electronic address: francois.houyengah@ch-dunkerque.fr.
Abstract
BACKGROUND & AIMS: Screening for malnutrition is not often done in the management of hospitalized patients. An original computer tool was developed at the Dunkerque hospital to detect readmissions of patients with malnutrition during a previous stay and generate e-mail alerts to the nutrition department. The aim of the study was to describe and evaluate this tool and the activity of the nutrition department in connection with the alerts sent. METHODS: The number of alerts sent, dietary consultations conducted, assessments of the nutritional status and the number of malnourished patients diagnosed were collected from September 1st to November 30th 2012. The positive predictive value (PPV) of the malnutrition screening tool was estimated. The evolution of the nutritional status between the last and the current hospitalization was also evaluated. RESULTS: A total of 531 e-mail alerts were sent (mean of 8.2 per working day), leading to 205 dietary consultations but only 144 recorded assessments of the nutritional status (lack of information in medical records). Of the latter, 128 diagnoses of malnutrition were made, i.e. a PPV of 88.9%, 95% Confidence Interval = [83.8%, 94%]. Overall, only one quarter of readmitted patients had improved nutritional status. CONCLUSIONS: The automatic e-mail alert system is operational and useful to effectively detect patients at risk of malnutrition and make follow up possible. In addition, an unfavorable evolution of the nutritional status of malnourished patients was observed.
BACKGROUND & AIMS: Screening for malnutrition is not often done in the management of hospitalized patients. An original computer tool was developed at the Dunkerque hospital to detect readmissions of patients with malnutrition during a previous stay and generate e-mail alerts to the nutrition department. The aim of the study was to describe and evaluate this tool and the activity of the nutrition department in connection with the alerts sent. METHODS: The number of alerts sent, dietary consultations conducted, assessments of the nutritional status and the number of malnourished patients diagnosed were collected from September 1st to November 30th 2012. The positive predictive value (PPV) of the malnutrition screening tool was estimated. The evolution of the nutritional status between the last and the current hospitalization was also evaluated. RESULTS: A total of 531 e-mail alerts were sent (mean of 8.2 per working day), leading to 205 dietary consultations but only 144 recorded assessments of the nutritional status (lack of information in medical records). Of the latter, 128 diagnoses of malnutrition were made, i.e. a PPV of 88.9%, 95% Confidence Interval = [83.8%, 94%]. Overall, only one quarter of readmitted patients had improved nutritional status. CONCLUSIONS: The automatic e-mail alert system is operational and useful to effectively detect patients at risk of malnutrition and make follow up possible. In addition, an unfavorable evolution of the nutritional status of malnourished patients was observed.