Background: The increase of aged patients with concurrent comorbities imply also a growth in end stage renal diseases incidence. HaemoDialysis (HD) has than a large medical, social and economical impact for health care systems. The personalization of the therapy is strategic to reduce costs and it can be performed basing on the information acquired by the collection the clinical data related to the treatment. The current work attempts to the extraction of an intradialytic hypotensive (IDH) events prediction index through the statistical mining of HD data coming from a multicentric study involving four different Italian and Swiss clinical centers. This work is part of the Project DialysIS, founded by a Cross-border Cooperation Programme (INTERREG IT/CH 2007–2013). Methods: Data referred to a total of 516 sessions performed on 70 adult patients undergoing dialysis treatment were collected. Clinical prescriptions, hydration status, dialysis machine data and hematochemical data were recorded in a flexible structured database. A statistical analysis was performed to find risk factors for IDH onset. The enrolled patients were classified in IDH prone and resistant, defining Hypotension Prone (HP), a patient who suffered of IDH in 2 or more session and Hypotension Resistant (HR) a patient who suffered at most 1 IDH episode. F and T test was performed on data to determine the significantly different parameters among the two groups. The new index J was defined as a weighted patient-specific combination these parameters, and was calculated for each session of each patient. Results: Using these patient specific coefficients, J results able to predict the 100% of treatments characterized by IDH events, with 38% of false positives (session at risk of IDH, without IDH onset). Conclusions: The J index can point out the risk to develop cardiovascular instabilities during each single treatment based on longitudinal observations of the patient specific parameters.
Statistical analysis of large amount of data aimed at the development of a index to predict intra-dialysis hypotensive events
VITO, DOMENICO;CASAGRANDE, GIUSTINA;BIANCHI, CAMILLA;COSTANTINO, MARIA LAURA
2015-01-01
Abstract
Background: The increase of aged patients with concurrent comorbities imply also a growth in end stage renal diseases incidence. HaemoDialysis (HD) has than a large medical, social and economical impact for health care systems. The personalization of the therapy is strategic to reduce costs and it can be performed basing on the information acquired by the collection the clinical data related to the treatment. The current work attempts to the extraction of an intradialytic hypotensive (IDH) events prediction index through the statistical mining of HD data coming from a multicentric study involving four different Italian and Swiss clinical centers. This work is part of the Project DialysIS, founded by a Cross-border Cooperation Programme (INTERREG IT/CH 2007–2013). Methods: Data referred to a total of 516 sessions performed on 70 adult patients undergoing dialysis treatment were collected. Clinical prescriptions, hydration status, dialysis machine data and hematochemical data were recorded in a flexible structured database. A statistical analysis was performed to find risk factors for IDH onset. The enrolled patients were classified in IDH prone and resistant, defining Hypotension Prone (HP), a patient who suffered of IDH in 2 or more session and Hypotension Resistant (HR) a patient who suffered at most 1 IDH episode. F and T test was performed on data to determine the significantly different parameters among the two groups. The new index J was defined as a weighted patient-specific combination these parameters, and was calculated for each session of each patient. Results: Using these patient specific coefficients, J results able to predict the 100% of treatments characterized by IDH events, with 38% of false positives (session at risk of IDH, without IDH onset). Conclusions: The J index can point out the risk to develop cardiovascular instabilities during each single treatment based on longitudinal observations of the patient specific parameters.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
