Literature DB >> 25954417

Pediatric readmission classification using stacked regularized logistic regression models.

Gregor Stiglic1, Fei Wang2, Adam Davey3, Zoran Obradovic3.   

Abstract

BACKGROUND: Regulations and privacy concerns often hinder exchange of healthcare data between hospitals or other healthcare providers. Sharing predictive models built on original data and averaging their results offers an alternative to more efficient prediction of outcomes on new cases. Although one can choose from many techniques to combine outputs from different predictive models, it is difficult to find studies that try to interpret the results obtained from ensemble-learning methods.
METHODS: We propose a novel approach to classification based on models from different hospitals that allows a high level of performance along with comprehensibility of obtained results. Our approach is based on regularized sparse regression models in two hierarchical levels and exploits the interpretability of obtained regression coefficients to rank the contribution of hospitals in terms of outcome prediction.
RESULTS: The proposed approach was used to predict the 30-days all-cause readmissions for pediatric patients in 54 Californian hospitals. Using repeated holdout evaluation, including more than 60,000 hospital discharge records, we compared the proposed approach to alternative approaches. The performance of two-level classification model was measured using the Area Under the ROC Curve (AUC) with an additional evaluation that uncovered the importance and contribution of each single data source (i.e. hospital) to the final result. The results for the best distributed model (AUC=0.787, 95% CI: 0.780-0.794) demonstrate no significant difference in terms of AUC performance when compared to a single elastic net model built on all available data (AUC=0.789, 95% CI: 0.781-0.796).
CONCLUSIONS: This paper presents a novel approach to improved classification with shared predictive models for environments where centralized collection of data is not possible. The significant improvements in classification performance and interpretability of results demonstrate the effectiveness of our approach.

Entities:  

Mesh:

Year:  2014        PMID: 25954417      PMCID: PMC4419883     

Source DB:  PubMed          Journal:  AMIA Annu Symp Proc        ISSN: 1559-4076


  16 in total

1.  Combining electronic healthcare databases in Europe to allow for large-scale drug safety monitoring: the EU-ADR Project.

Authors:  Preciosa M Coloma; Martijn J Schuemie; Gianluca Trifirò; Rosa Gini; Ron Herings; Julia Hippisley-Cox; Giampiero Mazzaglia; Carlo Giaquinto; Giovanni Corrao; Lars Pedersen; Johan van der Lei; Miriam Sturkenboom
Journal:  Pharmacoepidemiol Drug Saf       Date:  2010-11-08       Impact factor: 2.890

2.  Hospital readmission: quality indicator or statistical inevitability?

Authors:  Brian K Alverson; James O'Callaghan
Journal:  Pediatrics       Date:  2013-08-26       Impact factor: 7.124

3.  Measuring hospital quality using pediatric readmission and revisit rates.

Authors:  Naomi S Bardach; Eric Vittinghoff; Renée Asteria-Peñaloza; Jeffrey D Edwards; Jinoos Yazdany; Henry C Lee; W John Boscardin; Michael D Cabana; R Adams Dudley
Journal:  Pediatrics       Date:  2013-08-26       Impact factor: 7.124

4.  A study in transfer learning: leveraging data from multiple hospitals to enhance hospital-specific predictions.

Authors:  Jenna Wiens; John Guttag; Eric Horvitz
Journal:  J Am Med Inform Assoc       Date:  2014-01-30       Impact factor: 4.497

5.  Development, validation, and results of a measure of 30-day readmission following hospitalization for pneumonia.

Authors:  Peter K Lindenauer; Sharon-Lise T Normand; Elizabeth E Drye; Zhenqiu Lin; Katherine Goodrich; Mayur M Desai; Dale W Bratzler; Walter J O'Donnell; Mark L Metersky; Harlan M Krumholz
Journal:  J Hosp Med       Date:  2011-01-05       Impact factor: 2.960

6.  An administrative claims measure suitable for profiling hospital performance based on 30-day all-cause readmission rates among patients with acute myocardial infarction.

Authors:  Harlan M Krumholz; Zhenqiu Lin; Elizabeth E Drye; Mayur M Desai; Lein F Han; Michael T Rapp; Jennifer A Mattera; Sharon-Lise T Normand
Journal:  Circ Cardiovasc Qual Outcomes       Date:  2011-03

7.  Hospital utilization and characteristics of patients experiencing recurrent readmissions within children's hospitals.

Authors:  Jay G Berry; David E Hall; Dennis Z Kuo; Eyal Cohen; Rishi Agrawal; Chris Feudtner; Matt Hall; Jacqueline Kueser; William Kaplan; John Neff
Journal:  JAMA       Date:  2011-02-16       Impact factor: 56.272

8.  Pediatric readmission prevalence and variability across hospitals.

Authors:  Jay G Berry; Sara L Toomey; Alan M Zaslavsky; Ashish K Jha; Mari M Nakamura; David J Klein; Jeremy Y Feng; Shanna Shulman; Vincent W Chiang; Vincent K Chiang; William Kaplan; Matt Hall; Mark A Schuster
Journal:  JAMA       Date:  2013-01-23       Impact factor: 56.272

9.  EXpectation Propagation LOgistic REgRession (EXPLORER): distributed privacy-preserving online model learning.

Authors:  Shuang Wang; Xiaoqian Jiang; Yuan Wu; Lijuan Cui; Samuel Cheng; Lucila Ohno-Machado
Journal:  J Biomed Inform       Date:  2013-04-04       Impact factor: 6.317

10.  Combining knowledge and data driven insights for identifying risk factors using electronic health records.

Authors:  Jimeng Sun; Jianying Hu; Dijun Luo; Marianthi Markatou; Fei Wang; Shahram Edabollahi; Steven E Steinhubl; Zahra Daar; Walter F Stewart
Journal:  AMIA Annu Symp Proc       Date:  2012-11-03
View more
  3 in total

1.  Applicability of predictive models for 30-day unplanned hospital readmission risk in paediatrics: a systematic review.

Authors:  Ines Marina Niehaus; Nina Kansy; Stephanie Stock; Jörg Dötsch; Dirk Müller
Journal:  BMJ Open       Date:  2022-03-30       Impact factor: 2.692

2.  Comprehensible Predictive Modeling Using Regularized Logistic Regression and Comorbidity Based Features.

Authors:  Gregor Stiglic; Petra Povalej Brzan; Nino Fijacko; Fei Wang; Boris Delibasic; Alexandros Kalousis; Zoran Obradovic
Journal:  PLoS One       Date:  2015-12-08       Impact factor: 3.240

3.  Are Randomized Controlled Trials the (G)old Standard? From Clinical Intelligence to Prescriptive Analytics.

Authors:  Sven Van Poucke; Michiel Thomeer; John Heath; Milan Vukicevic
Journal:  J Med Internet Res       Date:  2016-07-06       Impact factor: 5.428

  3 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.