2
views
0
recommends
+1 Recommend
0 collections
    0
    shares
      • Record: found
      • Abstract: found
      • Article: found
      Is Open Access

      Derivation of an electronic frailty index for predicting short‐term mortality in heart failure: a machine learning approach

      research-article

      Read this article at

      Bookmark
          There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

          Abstract

          Aims

          Frailty may be found in heart failure patients especially in the elderly and is associated with a poor prognosis. However, assessment of frailty status is time‐consuming, and the electronic frailty indices developed using health records have served as useful surrogates. We hypothesized that an electronic frailty index developed using machine learning can improve short‐term mortality prediction in patients with heart failure.

          Methods and results

          This was a retrospective observational study that included patients admitted to nine public hospitals for heart failure from Hong Kong between 2013 and 2017. Age, sex, variables in the modified frailty index, Deyo's Charlson co‐morbidity index ( 2), neutrophil‐to‐lymphocyte ratio (NLR), and prognostic nutritional index at baseline were analysed. Gradient boosting, which is a supervised sequential ensemble learning algorithm with weak prediction submodels (typically decision trees), was applied to predict mortality. Variables were ranked in the order of importance with a total score of 100 and used to build the frailty models. Comparisons were made with decision tree and multivariable logistic regression. A total of 8893 patients (median: age 81, Q1–Q3: 71–87 years old) were included, in whom 9% had 30 day mortality and 17% had 90 day mortality. Prognostic nutritional index, age, and NLR were the most important variables predicting 30 day mortality (importance score: 37.4, 32.1, and 20.5, respectively) and 90 day mortality (importance score: 35.3, 36.3, and 14.6, respectively). Gradient boosting significantly outperformed decision tree and multivariable logistic regression. The area under the curve from a five‐fold cross validation was 0.90 for gradient boosting and 0.87 and 0.86 for decision tree and logistic regression in predicting 30 day mortality. For the prediction of 90 day mortality, the area under the curve was 0.92, 0.89, and 0.86 for gradient boosting, decision tree, and logistic regression, respectively.

          Conclusions

          The electronic frailty index based on co‐morbidities, inflammation, and nutrition information can readily predict mortality outcomes. Their predictive performances were significantly improved by gradient boosting techniques.

          Related collections

          Most cited references57

          • Record: found
          • Abstract: found
          • Article: found

          Frailty in elderly people

          Frailty is the most problematic expression of population ageing. It is a state of vulnerability to poor resolution of homoeostasis after a stressor event and is a consequence of cumulative decline in many physiological systems during a lifetime. This cumulative decline depletes homoeostatic reserves until minor stressor events trigger disproportionate changes in health status. In landmark studies, investigators have developed valid models of frailty and these models have allowed epidemiological investigations that show the association between frailty and adverse health outcomes. We need to develop more efficient methods to detect frailty and measure its severity in routine clinical practice, especially methods that are useful for primary care. Such progress would greatly inform the appropriate selection of elderly people for invasive procedures or drug treatments and would be the basis for a shift in the care of frail elderly people towards more appropriate goal-directed care. Copyright © 2013 Elsevier Ltd. All rights reserved.
            Bookmark
            • Record: found
            • Abstract: found
            • Article: not found

            Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases

            R Deyo (1992)
            Administrative databases are increasingly used for studying outcomes of medical care. Valid inferences from such data require the ability to account for disease severity and comorbid conditions. We adapted a clinical comorbidity index, designed for use with medical records, for research relying on International Classification of Diseases (ICD-9-CM) diagnosis and procedure codes. The association of this adapted index with health outcomes and resource use was then examined with a sample of Medicare beneficiaries who underwent lumbar spine surgery in 1985 (n = 27,111). The index was associated in the expected direction with postoperative complications, mortality, blood transfusion, discharge to nursing home, length of hospital stay, and hospital charges. These associations were observed whether the index incorporated data from multiple hospitalizations over a year's time, or just from the index surgical admission. They also persisted after controlling for patient age. We conclude that the adapted comorbidity index will be useful in studies of disease outcome and resource use employing administrative databases.
              Bookmark
              • Record: found
              • Abstract: found
              • Article: not found

              Frailty in older adults: evidence for a phenotype.

              Frailty is considered highly prevalent in old age and to confer high risk for falls, disability, hospitalization, and mortality. Frailty has been considered synonymous with disability, comorbidity, and other characteristics, but it is recognized that it may have a biologic basis and be a distinct clinical syndrome. A standardized definition has not yet been established. To develop and operationalize a phenotype of frailty in older adults and assess concurrent and predictive validity, the study used data from the Cardiovascular Health Study. Participants were 5,317 men and women 65 years and older (4,735 from an original cohort recruited in 1989-90 and 582 from an African American cohort recruited in 1992-93). Both cohorts received almost identical baseline evaluations and 7 and 4 years of follow-up, respectively, with annual examinations and surveillance for outcomes including incident disease, hospitalization, falls, disability, and mortality. Frailty was defined as a clinical syndrome in which three or more of the following criteria were present: unintentional weight loss (10 lbs in past year), self-reported exhaustion, weakness (grip strength), slow walking speed, and low physical activity. The overall prevalence of frailty in this community-dwelling population was 6.9%; it increased with age and was greater in women than men. Four-year incidence was 7.2%. Frailty was associated with being African American, having lower education and income, poorer health, and having higher rates of comorbid chronic diseases and disability. There was overlap, but not concordance, in the cooccurrence of frailty, comorbidity, and disability. This frailty phenotype was independently predictive (over 3 years) of incident falls, worsening mobility or ADL disability, hospitalization, and death, with hazard ratios ranging from 1.82 to 4.46, unadjusted, and 1.29-2.24, adjusted for a number of health, disease, and social characteristics predictive of 5-year mortality. Intermediate frailty status, as indicated by the presence of one or two criteria, showed intermediate risk of these outcomes as well as increased risk of becoming frail over 3-4 years of follow-up (odds ratios for incident frailty = 4.51 unadjusted and 2.63 adjusted for covariates, compared to those with no frailty criteria at baseline). This study provides a potential standardized definition for frailty in community-dwelling older adults and offers concurrent and predictive validity for the definition. It also finds that there is an intermediate stage identifying those at high risk of frailty. Finally, it provides evidence that frailty is not synonymous with either comorbidity or disability, but comorbidity is an etiologic risk factor for, and disability is an outcome of, frailty. This provides a potential basis for clinical assessment for those who are frail or at risk, and for future research to develop interventions for frailty based on a standardized ascertainment of frailty.
                Bookmark

                Author and article information

                Contributors
                qingpeng.zhang@cityu.edu.hk
                g.tse@surrey.ac.uk
                Journal
                ESC Heart Fail
                ESC Heart Fail
                10.1002/(ISSN)2055-5822
                EHF2
                ESC Heart Failure
                John Wiley and Sons Inc. (Hoboken )
                2055-5822
                03 June 2021
                August 2021
                : 8
                : 4 ( doiID: 10.1002/ehf2.v8.4 )
                : 2837-2845
                Affiliations
                [ 1 ] Research Department of Practice and Policy, School of Pharmacy University College London London UK
                [ 2 ] School of Data Science City University of Hong Kong Hong Kong SAR China
                [ 3 ] Cardiovascular Analytics Group, Laboratory of Cardiovascular Physiology, LKS Institute of Health Sciences Chinese University of Hong Kong Hong Kong SAR China
                [ 4 ] Faculty of Arts and Science University of Toronto Toronto Ontario Canada
                [ 5 ] Tianjin Key Laboratory of Ionic‐Molecular Function of Cardiovascular Disease, Department of Cardiology, Tianjin Institute of Cardiology Second Hospital of Tianjin Medical University Tianjin China
                [ 6 ] Second Department of Cardiology Evangelismos General Hospital Athens Greece
                [ 7 ] Faculty of Health and Medical Sciences University of Surrey Guildford UK
                [ 8 ] Centre for Safe Medication Practice and Research, Department of Pharmacology and Pharmacy The University of Hong Kong Hong Kong SAR China
                Author notes
                [*] [* ] Correspondence to: Qingpeng Zhang, School of Data Science, City University of Hong Kong, Hong Kong SAR, China. Email: qingpeng.zhang@ 123456cityu.edu.hk

                Gary Tse, Faculty of Health and Medical Sciences, University of Surrey, Guildford, UK and Tianjin Key Laboratory of Ionic‐Molecular Function of Cardiovascular Disease, Department of Cardiology, Tianjin Institute of Cardiology, Second Hospital of Tianjin Medical University, Tianjin 300211, China. Email: g.tse@ 123456surrey.ac.uk

                [ † ]

                Chengsheng Ju and Jiandong Zhou for joint first authorship.

                Article
                EHF213358 ESCHF-20-00937
                10.1002/ehf2.13358
                8318426
                34080784
                27c3cab5-2f7f-4d92-b8e0-c6e80a1d8bf5
                © 2021 The Authors. ESC Heart Failure published by John Wiley & Sons Ltd on behalf of the European Society of Cardiology.

                This is an open access article under the terms of the http://creativecommons.org/licenses/by/4.0/ License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

                Page count
                Figures: 3, Tables: 2, Pages: 9, Words: 3781
                Product
                Categories
                Original Research Article
                Original Research Articles
                Custom metadata
                2.0
                August 2021
                Converter:WILEY_ML3GV2_TO_JATSPMC version:6.0.4 mode:remove_FC converted:28.07.2021

                frailty index,heart failure,mortality,inflammation,nutrition,machine learning

                Comments

                Comment on this article