The University of Southampton
University of Southampton Institutional Repository

Five years survival on hemodialysis predicted by artificial neural network model

Five years survival on hemodialysis predicted by artificial neural network model
Five years survival on hemodialysis predicted by artificial neural network model
Introduction: Maintenance hemodialysis (HD) patients’ morbidity and mortality remain unacceptably high. It is important to identify risk factors affecting outcome and define their relative contribution.
Methods: The data of 93 patients who started HD at the Sheffield Kidney Institute between January 1997 and June 1998 were analyzed retrospectively. Potential risk predictors of 5-year survival were evaluated both at baseline and at 1-year follow-up, including changes from baseline.
Results: Patients median age was 60 years (range: 27-80), 75% were males and 23% were diabetics. According to the Index of Co-Existing Disease (ICED) score, 18% of patients were level 0-1 (low comorbidity), 56% level 2 and 26% level 3. The 5-year survival was 58%. The Cox models identified eight independent time-adjusted risk factors for 5-year survival, the importance of which as independent predictors was confirmed by logistic regression models. According to the artificial neural network (ANN) models, the relative importance of these factors was as follows: age (26.6%), baseline systolic blood pressure (24.1%), mean Kt/V during the first year (19.7%), baseline ICED score (14.8%), baseline diastolic blood pressure (8.6%), serum calcium change from baseline (2.8%), blood urea change from baseline (2.1%), serum creatinine change from baseline (1.3%). This ANN model had a high level of predictive performance as assessed by accuracy (93.3%) and by Receiver Operating Characteristic (ROC) curve analysis (AUC=0.92).
Conclusion: Integrated use of regression analysis and probabilistic models allow computation of individual risk of mortality in HD. This may help in optimizing care and costs.
hemodialysis, risk factors, iced, artificial neural network
1858-554X
10-19
Ahmed, B.
b3de9c41-9407-4fea-ad87-4c587554fdf3
Dimitrov, B.
366d715f-ffd9-45a1-8415-65de5488472f
Perna, A.
a4b182ec-c177-4ba4-bfe3-a8472182b083
Remuzzi, G.
c2b8dcb0-a5e4-4f29-ad90-ab77ce71e207
Nahas, M.
3f55d2d3-6ba5-45df-b61c-3c49f00445af
Ahmed, B.
b3de9c41-9407-4fea-ad87-4c587554fdf3
Dimitrov, B.
366d715f-ffd9-45a1-8415-65de5488472f
Perna, A.
a4b182ec-c177-4ba4-bfe3-a8472182b083
Remuzzi, G.
c2b8dcb0-a5e4-4f29-ad90-ab77ce71e207
Nahas, M.
3f55d2d3-6ba5-45df-b61c-3c49f00445af

Ahmed, B., Dimitrov, B., Perna, A., Remuzzi, G. and Nahas, M. (2009) Five years survival on hemodialysis predicted by artificial neural network model. Arab Journal of Nephrology and Transplantation, 2 (1), 10-19. (doi:10.4314/ajnt.v2i1.58838).

Record type: Article

Abstract

Introduction: Maintenance hemodialysis (HD) patients’ morbidity and mortality remain unacceptably high. It is important to identify risk factors affecting outcome and define their relative contribution.
Methods: The data of 93 patients who started HD at the Sheffield Kidney Institute between January 1997 and June 1998 were analyzed retrospectively. Potential risk predictors of 5-year survival were evaluated both at baseline and at 1-year follow-up, including changes from baseline.
Results: Patients median age was 60 years (range: 27-80), 75% were males and 23% were diabetics. According to the Index of Co-Existing Disease (ICED) score, 18% of patients were level 0-1 (low comorbidity), 56% level 2 and 26% level 3. The 5-year survival was 58%. The Cox models identified eight independent time-adjusted risk factors for 5-year survival, the importance of which as independent predictors was confirmed by logistic regression models. According to the artificial neural network (ANN) models, the relative importance of these factors was as follows: age (26.6%), baseline systolic blood pressure (24.1%), mean Kt/V during the first year (19.7%), baseline ICED score (14.8%), baseline diastolic blood pressure (8.6%), serum calcium change from baseline (2.8%), blood urea change from baseline (2.1%), serum creatinine change from baseline (1.3%). This ANN model had a high level of predictive performance as assessed by accuracy (93.3%) and by Receiver Operating Characteristic (ROC) curve analysis (AUC=0.92).
Conclusion: Integrated use of regression analysis and probabilistic models allow computation of individual risk of mortality in HD. This may help in optimizing care and costs.

Full text not available from this repository.

More information

Published date: 2009
Keywords: hemodialysis, risk factors, iced, artificial neural network
Organisations: Primary Care & Population Sciences

Identifiers

Local EPrints ID: 365792
URI: https://eprints.soton.ac.uk/id/eprint/365792
ISSN: 1858-554X
PURE UUID: 960ffa4d-3581-49d9-a684-2b0665e1deca

Catalogue record

Date deposited: 16 Jun 2014 12:04
Last modified: 16 Jul 2019 21:03

Export record

Altmetrics

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of https://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×