Transforming Mortality Prediction: A Transformer-Based Mortality Prediction Model.
Objectives Mortality prediction and the identification of mortality risks are central to social and biological sciences. Traditional models often assess linear associations between single risk factors and mortality. Transformer models, capable of capturing long-term dependencies across multiple vari...
| Publicado en: | Journals of Gerontology Series B: Psychological Sciences & Social Sciences Vol. 80; no. 7; pp. 1 - 10 |
|---|---|
| Autores principales: | , , , |
| Formato: | Artículo |
| Publicado: |
Oxford University Press / USA
Jul2025
|
| Materias: | |
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ssf&AN=189082057&site=ehost-live header: @attributes: shortDbName: ssf uiTerm: 189082057 longDbName: Social Sciences Full Text (H.W. Wilson) uiTag: AN controlInfo: bkinfo: jinfo: jid: 10795014 JGB jtl: Journals of Gerontology Series B: Psychological Sciences & Social Sciences issn: 10795014 maglogo: N pubinfo: dt: Jul2025 vid: 80 iid: 7 pid: 622 pub: Oxford University Press / USA artinfo: ui: 189082057 10.1093/geronb/gbaf089 ppf: 1 ppct: 9 formats: tig: atl: Transforming Mortality Prediction: A Transformer-Based Mortality Prediction Model. aug: au: Weiss, Jordan Azhir, Alaleh Ram, Nilam Rehkopf, David H affil: Stanford Center on Longevity, Stanford University, Stanford, California, USA Optimal Aging Institute & Division of Precision Medicine, NYU Grossman School of Medicine, New York City, New York, USA Department of Medicine, Brigham and Women's Hospital, Boston, Massachusetts, USA Department of Communication, Stanford University, Palo Alto, California, USA Department of Psychology, Stanford University, Palo Alto, California, USA Department of Epidemiology and Population Health, Stanford University, Palo Alto, California, USA Division of Primary Care and Population Health, Department of Medicine, Stanford University School of Medicine, Stanford University, Stanford, California, USA su: United States Self-evaluation Health status indicators Satisfaction Interviewing Population health Causes of death Sociodemographic factors Stroke Affect (Psychology) Mental depression Diabetes Neuroses Mortality risk factors Risk assessment Random forest algorithms Prediction models Secondary analysis Receiver operating characteristic curves Multiple regression analysis Descriptive statistics Artificial neural networks Life course approach Machine learning Decision trees Data analysis software Proportional hazards models Algorithms sug: subj: Self-evaluation Health status indicators Satisfaction Interviewing Population health Causes of death Sociodemographic factors Stroke Affect (Psychology) Mental depression Diabetes Neuroses United States Mortality risk factors Risk assessment Random forest algorithms Prediction models Secondary analysis Receiver operating characteristic curves Multiple regression analysis Descriptive statistics Artificial neural networks Life course approach Machine learning Decision trees Data analysis software Proportional hazards models Algorithms keyword: Large language models Life course Large language models Life course ab: Objectives Mortality prediction and the identification of mortality risks are central to social and biological sciences. Traditional models often assess linear associations between single risk factors and mortality. Transformer models, capable of capturing long-term dependencies across multiple variables, offer a novel approach to mortality prediction. This study introduces a transformer-based model applied to data from the Health and Retirement Study (HRS). Methods We analyzed data provided by 38,193 adults aged ≥ 50 years participating in the HRS, a longitudinal U.S. study surveyed biennially since 1992. Linked mortality data were obtained from the National Death Index and postmortem interviews. Using the transformer architecture, we modeled changes in 126 risk factors spanning financial, physical, and mental health domains manifesting over 29 years. Prediction performance was assessed across multiple settings, with traditional statistical and machine learning (ML) models serving as benchmarks. Results Over a median follow-up of 9 years, 17,448 deaths occurred (crude rate: 39.6 per 1,000 person-years). The transformer model consistently outperformed traditional and ML methods, achieving a 2-fold improvement in average precision scores for next-wave mortality prediction relative to the best benchmark model. Discussion Transformer-based models, such as BEHRT, significantly enhance mortality prediction compared with traditional approaches. These findings highlight the potential of transformer neural network models in social science-focused population health research on aging. pubtype: Academic Journal doctype: Article src: R language: English refInfo: copyright: @attributes: flag: N holdings: @attributes: islocal: N |
|---|