CoChat
Sign Up
Cited by 73Machine Learning in HealthcareArtificial Intelligence in Healthcare and EducationArtificial Intelligence in Healthcare

Integrating machine learning and artificial intelligence in life-course epidemiology: pathways to innovative public health solutions

Open access
Chen, Shanquan, Yu, Jiazhou, Chamouni, Sarah, Wang, Yuqi, Li, Yunfei·2024·BMC Medicine
DOI 10.1186/s12916-024-03566-xPMC PMC11367811
Open this paper in CoChat — free accountOpen PDF

Abstract

The integration of machine learning (ML) and artificial intelligence (AI) techniques in life-course epidemiology offers remarkable opportunities to advance our understanding of the complex interplay between biological, social, and environmental factors that shape health trajectories across the lifespan. This perspective summarizes the current applications, discusses future potential and challenges, and provides recommendations for harnessing ML and AI technologies to develop innovative public health solutions. ML and AI have been increasingly applied in epidemiological studies, demonstrating their ability to handle large, complex datasets, identify intricate patterns and associations, integrate multiple and multimodal data types, improve predictive accuracy, and enhance causal inference methods. In life-course epidemiology, these techniques can help identify sensitive periods and critical windows for intervention, model complex interactions between risk factors, predict individual and population-level disease risk trajectories, and strengthen causal inference in observational studies. By leveraging the five principles of life-course research proposed by Elder and Shanahan—lifespan development, agency, time and place, timing, and linked lives—we discuss a framework for applying ML and AI to uncover novel insights and inform targeted interventions. However, the successful integration of these technologies faces challenges related to data quality, model interpretability, bias, privacy, and equity. To fully realize the potential of ML and AI in life-course epidemiology, fostering interdisciplinary collaborations, developing standardized guidelines, advocating for their integration in public health decision-making, prioritizing fairness, and investing in training and capacity building are essential. By responsibly harnessing the power of ML and AI, we can take significant steps towards creating healthier and more equitable futures across the life course.

AI snapshot

Written from the paper’s full text.

TL;DR

This perspective argues that machine learning and artificial intelligence can advance life-course epidemiology by identifying sensitive periods, modeling complex interactions, predicting disease trajectories, and strengthening causal inference, framed around Elder and Shanahan's five life-course principles. It also catalogues substantial challenges—data quality, interpretability, bias, privacy, equity, and possible overreliance on ML/AI—and offers recommendations for responsible integration into public health.

Key findings

  • ML/AI applications in epidemiology already show concrete results: a deep learning study used 216,152 retinal photographs from South Korea, Singapore, and the UK to derive coronary artery calcium scores comparable to CT scans; Ward et al. (2020) trained models on EHRs of 262,923 individuals with performance comparable to or better than traditional atherosclerotic cardiovascular disease risk equations; Bhagwat et al. (2018) predicted conversion from mild cognitive impairment to Alzheimer's disease with accuracy as high as 0.90; and Ye et al. (2018) validated a hypertension risk model on EHRs of more than 1.5 million individuals, later deployed in Maine.
  • The paper maps ML/AI techniques onto Elder and Shanahan's five principles: lifespan development (deep learning, longitudinal modeling), agency (decision trees, reinforcement learning), time and place (spatial modeling, time series, RNNs/LSTMs, Transformers/BERT), timing (survival forests, Bayesian networks), and linked lives (social network analysis, agent-based modeling, large language models).
  • It cautions against overreliance: a systematic review of 65 randomized controlled trials of AI-based prediction tools found approximately 40% failed to show statistically significant clinical benefit versus standard care, and the "Fragile Families Challenge" showed state-of-the-art ML models improved only marginally over simple benchmarks for predicting life outcomes.
  • Recommendations include interdisciplinary collaboration, standardized guidelines, integration into public health decision-making, prioritizing equity and fairness, and investment in training and capacity building.

Methods

Not stated in the paper (this is a perspective/narrative synthesis, not an empirical study). The paper summarizes current applications, opportunities, challenges, and recommendations, drawing on cited examples across cardiovascular disease, cancer, neurodegenerative disease, infectious disease, environmental exposure, and social determinants of health. No systematic search strategy, inclusion criteria, or primary data analysis is described.

Limitations

The paper explicitly states it is not a systematic review, that the information presented may be subject to selection bias, and that its narrative approach may not capture all relevant studies or viewpoints, with cited examples selected for relevance and impact rather than representing an exhaustive body of evidence. Evident from the text: the paper presents no original data or empirical validation of its proposed framework, and the framework's application to the five life-course principles is illustrative rather than tested.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1186/s12916-024-03566-x

Cites

69 of 69 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 97 cited sources checked · 0 supported · 0 flagged

  • Interpretationcitation

    Applying the five life-course principles can help researchers identify sensitive periods and critical windows during which interventions may be most effective in preventing or mitigating the risk of chronic diseases.

    “By applying these principles, researchers can identify sensitive periods and critical windows during which interventions may be most effective in preventing or mitigating the risk of chronic diseases.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [2] Elder Jr GH, Shanahan MJ. The Life Course and Human Development. In: Damon W, Lerner RM, editors. Handbook of Child Psychology. Volume 1, edn. New Jersey: Wiley; 2007.

      Not checked in the free run.

  • Premisecitation

    Bhagwat et al. (2018) used a neural-network algorithm to predict the conversion from mild cognitive impairment to Alzheimer's disease with as high as 0.90 accuracy.

    “Bhagwat et al. (2018) used a neural-network algorithm to predict the conversion from mild cognitive impairment to Alzheimer’s disease with as high as 0.90 accuracy [18].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [18] Bhagwat N, Viviano JD, Voineskos AN, Chakravarty MM, Alzheimer’s Disease Neuroimaging I. Modeling and prediction of clinical symptom trajectories in Alzheimer’s disease using longitudinal data. PLoS Comput Biol. 2018;14(9):e1006376. source

      Not checked in the free run.

  • Interpretationreasoning

    ML and AI techniques can develop predictive models that inform personalized interventions and public health strategies.

    “These techniques offer the ability to handle vast amounts of complex, high-dimensional data, identify intricate patterns and associations, and develop predictive models that can inform personalized interventions and public health strategies.”
    Quote in the paperSubstantial claim
  • Premisecitation

    By integrating neuroimaging, genetic, and clinical data, researchers have developed models that can identify individuals at high risk of developing Alzheimer's disease.

    “By integrating neuroimaging, genetic, and clinical data, researchers have developed models that can identify individuals at high risk of developing Alzheimer’s disease [15–18].”
    Quote in the paperNot a substantial claim0 of 4 sources checked
    • cites [15] Grueso S, Viejo-Sobera R. Machine learning methods for predicting progression from mild cognitive impairment to Alzheimer’s disease dementia: a systematic review. Alzheimers Res Ther. 2021;13(1):162. source

      Not checked in the free run.

    • cites [16] Tang AS, Rankin KP, Cerono G, Miramontes S, Mills H, Roger J, Zeng B, Nelson C, Soman K, Woldemariam S, et al. Leveraging electronic health records and knowledge networks for Alzheimer’s disease prediction and sex-specific biological insights. Nat Aging. 2024;4(3):379–95. source

      Not checked in the free run.

    • cites [17] Gao XR, Chiariglione M, Qin K, Nuytemans K, Scharre DW, Li YJ, Martin ER. Explainable machine learning aggregates polygenic risk scores and electronic health records for Alzheimer’s disease prediction. Sci Rep. 2023;13(1):450. source

      Not checked in the free run.

    • cites [18] Bhagwat N, Viviano JD, Voineskos AN, Chakravarty MM, Alzheimer’s Disease Neuroimaging I. Modeling and prediction of clinical symptom trajectories in Alzheimer’s disease using longitudinal data. PLoS Comput Biol. 2018;14(9):e1006376. source

      Not checked in the free run.

  • Premisecitation

    ML and AI techniques have been applied to predict cancer prognosis and estimate treatment response based on genomic and clinical data.

    “These techniques have been applied to predict cancer prognosis and estimate treatment response based on genomic and clinical data [8–14].”
    Quote in the paperNot a substantial claim0 of 7 sources checked
    • cites [8] Tran KA, Kondrashova O, Bradley A, Williams ED, Pearson JV, Waddell N. Deep learning in cancer diagnosis, prognosis and treatment selection. Genome Med. 2021;13(1):152. source

      Not checked in the free run.

    • cites [9] Waljee AK, Weinheimer-Haus EM, Abubakar A, Ngugi AK, Siwo GH, Kwakye G, Singal AG, Rao A, Saini SD, Read AJ, et al. Artificial intelligence and machine learning for early detection and diagnosis of colorectal cancer in sub-Saharan Africa. Gut. 2022;71(7):1259–65. source

      Not checked in the free run.

    • cites [10] Zhang B, Shi H, Wang H. Machine learning and AI in cancer prognosis, prediction, and treatment selection: a critical approach. J Multidiscip Healthc. 2023;16:1779–91. source

      Not checked in the free run.

    • cites [11] Lee EY, Maloney NJ, Cheng K, Bach DQ. Machine learning for precision dermatology: advances, opportunities, and outlook. J Am Acad Dermatol. 2021;84(5):1458–9. source

      Not checked in the free run.

    • cites [12] Swanson K, Wu E, Zhang A, Alizadeh AA, Zou J. From patterns to patients: advances in clinical machine learning for cancer diagnosis, prognosis, and treatment. Cell. 2023;186(8):1772–91. source

      Not checked in the free run.

    • cites [13] Lu L, Dercle L, Zhao B, Schwartz LH. Deep learning for the prediction of early on-treatment response in metastatic colorectal cancer from serial medical imaging. Nat Commun. 2021;12(1):6654. source

      Not checked in the free run.

    • cites [14] Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, Thrun S. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017;542(7639):115–8. source

      Not checked in the free run.

  • Classified as probably not a substantial claim. 33
  • May combine two claims. 28

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The life-course approach is based on five key principles: lifespan development, agency, time and place, timing, and linked lives.

    “As defined by Elder and Shanahan, the life-course approach is based on five key principles: lifespan development, agency, time and place, timing, and linked lives (Fig. 1) [2].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [2] Elder Jr GH, Shanahan MJ. The Life Course and Human Development. In: Damon W, Lerner RM, editors. Handbook of Child Psychology. Volume 1, edn. New Jersey: Wiley; 2007.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Premisereasoning

    ML and AI can integrate multiple data types, such as electronic health records, genomic data, and environmental exposures, to provide a more comprehensive understanding of the factors contributing to health outcomes across the life course.

    “ML and AI can integrate multiple data types, such as electronic health records (EHRs), genomic data, and environmental exposures, to provide a more comprehensive understanding of the factors contributing to health outcomes across the life course.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Lu et al. (2021) developed a deep learning network to predict early on-treatment response in metastatic colorectal cancer, which outperformed traditional methods.

    “For instance, Lu et al. (2021) developed a deep learning network to predict early on-treatment response in metastatic colorectal cancer, which outperformed traditional methods [13].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [13] Lu L, Dercle L, Zhao B, Schwartz LH. Deep learning for the prediction of early on-treatment response in metastatic colorectal cancer from serial medical imaging. Nat Commun. 2021;12(1):6654. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationreasoning

    The integration of ML and AI techniques in life-course epidemiology has the potential to revolutionize understanding of the complex determinants of diseases and inform the development of more targeted and effective public health interventions.

    “The integration of ML and AI techniques in life-course epidemiology has the potential to revolutionize our understanding of the complex determinants of diseases and inform the development of more targeted and effective public health interventions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Life-course epidemiology examines the long-term effects of biological, behavioral, and social exposures during gestation, childhood, adolescence, and adulthood on the development of chronic diseases later in life.

    “Life-course epidemiology is a field of study that examines the long-term effects of biological, behavioral, and social exposures during gestation, childhood, adolescence, and adulthood on the development of chronic diseases later in life [1].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [1] Wagner C, Carmeli C, Jackisch J, Kivimaki M, van der Linden BWA, Cullati S, Chiolero A. Life course epidemiology and public health. Lancet Public Health. 2024;9(4):e261–9. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    ML and AI techniques can handle vast amounts of complex, high-dimensional data.

    “These techniques offer the ability to handle vast amounts of complex, high-dimensional data, identify intricate patterns and associations, and develop predictive models that can inform personalized interventions and public health strategies.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    ML and AI have been applied to predict disease outbreaks and identify high-risk areas.

    “In the realm of infectious diseases, ML and AI have been applied to predict disease outbreaks and identify high-risk areas [19, 20].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [19] Dogan O, Tiwari S, Jabbar MA, Guggari S. A systematic review on AI/ML approaches against COVID-19 outbreak. Complex Intell Syst. 2021;7(5):2655–78. source

      Not checked in the free run.

    • cites [20] Bengtsson L, Gaudart J, Lu X, Moore S, Wetter E, Sallah K, Rebaudet S, Piarroux R. Using mobile phone data to predict the spatial spread of cholera. Sci Rep. 2015;5:8923. source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisenone given

    ML and AI have been increasingly applied in various areas of epidemiological studies, demonstrating their potential to advance understanding of health and diseases.

    “ML and AI have been increasingly applied in various areas of epidemiological studies, demonstrating their potential to advance our understanding of health and diseases.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    ML and AI approaches can enhance causal inference methods, allowing researchers to better estimate the effects of exposures on health outcomes in observational settings.

    “Furthermore, these approaches can enhance causal inference methods, allowing researchers to better estimate the effects of exposures on health outcomes in observational settings.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    A study using 216,152 retinal photographs from datasets in South Korea, Singapore, and the United Kingdom trained and validated deep learning algorithms, and retinal photograph-derived coronary artery calcium scores were comparable to those measured by CT scans.

    “For example, a study utilized 216,152 retinal photographs from datasets in South Korea, Singapore, and the United Kingdom to train and validate deep learning algorithms [6]. The retinal photograph-derived coronary artery calcium scores were found to be comparable to those measured by CT scans [6].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [6] Rim TH, Lee CJ, Tham YC, Cheung N, Yu M, Lee G, Kim Y, Ting DSW, Chong CCY, Choi YS, et al. Deep-learning-based cardiovascular risk stratification using coronary artery calcium scores predicted from retinal photographs. Lancet Digit Health. 2021;3(5):e306–16. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationnone given

    By leveraging ML and AI, researchers can uncover novel risk factors, identify critical windows for intervention, and predict individual disease trajectories with greater precision.

    “By leveraging the power of these innovative tools, researchers can uncover novel risk factors, identify critical windows for intervention, and predict individual disease trajectories with greater precision.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given

    Advanced ML and AI techniques can analyze multimodal data, including structured and unstructured text, audio, image, and video data.

    “Moreover, advanced ML and AI techniques can analyze multimodal data, including structured and unstructured text, audio, image, and video data, enabling the examination of diverse information sources such as MRI scans, X-rays, recordings of heartbeats or respiratory sounds, and physical activity and behavioral patterns.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Ward et al. (2020) used EHRs of 262,923 individuals to train and validate ML models, demonstrating performance comparable to or better than traditional equations for atherosclerotic cardiovascular disease risk.

    “Ward et al. (2020) used EHRs of 262,923 individuals to train and validate ML models, demonstrating performance that was comparable to or better than traditional equations for atherosclerotic cardiovascular disease risk [7].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [7] Ward A, Sarraju A, Chung S, Li J, Harrington R, Heidenreich P, Palaniappan L, Scheinker D, Rodriguez F. Machine learning and atherosclerotic cardiovascular disease risk prediction in a multi-ethnic population. NPJ Digit Med. 2020;3:125. source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Health and disease are influenced by the complex interplay of various factors across an individual's life span, and the timing and duration of exposures can have critical implications for future health outcomes.

    “This approach recognizes that health and disease are influenced by the complex interplay of various factors across an individual’s life span and that the timing and duration of these exposures can have critical implications for future health outcomes [1].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [1] Wagner C, Carmeli C, Jackisch J, Kivimaki M, van der Linden BWA, Cullati S, Chiolero A. Life course epidemiology and public health. Lancet Public Health. 2024;9(4):e261–9. source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisereasoning

    ML and AI techniques can identify intricate patterns and associations in data.

    “These techniques offer the ability to handle vast amounts of complex, high-dimensional data, identify intricate patterns and associations, and develop predictive models that can inform personalized interventions and public health strategies.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Chu et al. (2020) proposed an adversarial deep treatment effect prediction (ADTEP) model to predict treatment effects using heterogeneous EHR data.

    “Similarly, Chu et al. (2020) proposed an adversarial deep treatment effect prediction (ADTEP) model to predict treatment effects using heterogeneous EHR data [29].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [29] Chu J, Dong W, Wang J, He K, Huang Z. Treatment effect prediction with adversarial deep learning using electronic health records. BMC Med Inform Decis Mak. 2020;20(Suppl 4):139. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    ML and AI can uncover intricate, non-linear relationships between exposures and health outcomes that may not be easily identified using traditional statistical methods.

    “Identifying complex patterns: These techniques can uncover intricate, non-linear relationships between exposures and health outcomes, which may not be easily identified using traditional statistical methods.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    ML and AI techniques, such as spatial modeling and time series analysis, can analyze the effects of place and time on health outcomes and identify key contextual factors that may influence the effectiveness of public health interventions.

    “ML and AI techniques, such as spatial modeling and time series analysis, can analyze the effects of place and time on health outcomes and identify key contextual factors that may influence the effectiveness of public health interventions [39, 41].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [39] Credit K, Lehnert M. A structured comparison of causal machine learning methods to assess heterogeneous treatment effects in spatial data. In: Journal of Geographical Systems. 2023.

      Not checked in the free run.

    • cites [41] Huang C, Petukhina A. Modern Machine Learning Methods for Time Series Analysis. In: Applied Time Series Analysis and Forecasting with Python. Switzerland: Springer; 2022. p. 341–61. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    This information can guide the development of targeted interventions and policies to prevent and manage chronic diseases.

    “This information can guide the development of targeted interventions and policies to prevent and manage chronic diseases.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    ML and AI have been employed to investigate the social determinants of health and identify populations at high risk of adverse health outcomes.

    “Furthermore, ML and AI have been employed to investigate the social determinants of health and identify populations at high risk of adverse health outcomes [23].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [23] Guevara M, Chen S, Thomas S, Chaunzwa TL, Franco I, Kann BH, Moningi S, Qian JM, Goldstein M, Harper S, et al. Large language models to identify social determinants of health in electronic health records. NPJ Digit Med. 2024;7(1):6. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    ML and AI can process vast amounts of high-dimensional data, making them valuable tools for extracting meaningful insights from diverse data sources.

    “Handling large, complex datasets: ML and AI can process vast amounts of high-dimensional data, making them valuable tools for extracting meaningful insights from diverse data sources.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationcitation

    The five principles serve as a foundation for harnessing the potential of ML and AI to identify previously unknown risk factors, predict disease progression, and guide the development of targeted interventions.

    “These principles also serve as a foundation for harnessing the potential of ML and AI to identify previously unknown risk factors, predict disease progression, and guide the development of targeted interventions.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [2] Elder Jr GH, Shanahan MJ. The Life Course and Human Development. In: Damon W, Lerner RM, editors. Handbook of Child Psychology. Volume 1, edn. New Jersey: Wiley; 2007.

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The Hinson et al. tool was rapidly integrated into clinical practice within the Johns Hopkins Health System, contributing to more consistent and reliable disposition decisions and improved bed allocation during the COVID-19 pandemic.

    “This tool was rapidly integrated into clinical practice to support care decisions within the Johns Hopkins Health System, contributing to more consistent and reliable disposition decision and improved bed allocation during the pandemic [26].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [26] Hamilton AJ, Strauss AT, Martinez DA, Hinson JS, Levin S, Lin G, Klein EY. Machine learning and artificial intelligence: applications in healthcare epidemiology. Antimicrob Steward Healthc Epidemiol. 2021;1(1):e28. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    In life-course epidemiology, ML and AI offer opportunities by enabling researchers to identify sensitive periods, model complex interactions, predict disease risk trajectories, and enhance causal inference methods.

    “In life-course epidemiology that considers long-term effects of biological, behavioral, and social exposures during gestation, childhood, adolescence, and adulthood, ML and AI offer numerous opportunities by enabling researchers to identify sensitive periods, model complex interactions, predict disease risk trajectories, and enhance causal inference methods.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisecitation

    The five principles outlined by Elder and Shanahan offer a robust conceptual framework for comprehending the intricate and ever-changing aspects of health and disease throughout an individual's life course.

    “The five principles outlined by Elder and Shanahan offer a robust conceptual framework for comprehending the intricate and ever-changing aspects of health and disease throughout an individual’s life course [2].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [2] Elder Jr GH, Shanahan MJ. The Life Course and Human Development. In: Damon W, Lerner RM, editors. Handbook of Child Psychology. Volume 1, edn. New Jersey: Wiley; 2007.

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    ML and AI can contribute to causal inference by identifying potential causal pathways and controlling for confounding factors in observational studies.

    “In the end, ML and AI can also contribute to causal inference by identifying potential causal pathways and controlling for confounding factors in observational studies.”
    • Also stated in: § Current applications of ML and AI in epidemiology · plain
      “Enhancing causal inference: While primarily used for prediction, ML and AI can also contribute to causal inference by identifying potential causal pathways and controlling for confounding factors in observational studies.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    ML and AI approaches often achieve higher predictive accuracy than traditional methods, particularly when dealing with complex datasets.

    “Improving predictive accuracy: These approaches often achieve higher predictive accuracy than traditional methods, particularly when dealing with complex datasets, enabling the development of more precise risk prediction models.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    ML and AI techniques, such as survival analysis (e.g. Survival Forest) and Bayesian networks, can model the time-varying effects of exposures on health outcomes and identify optimal timing and targeting of interventions.

    “ML and AI techniques, such as survival analysis (e.g. Survival Forest) and Bayesian networks, can model the time-varying effects of exposures on health outcomes and identify optimal timing and targeting of interventions [37, 38].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [37] Jawadekar N, Kezios K, Odden MC, Stingone JA, Calonico S, Rudolph K, Zeki Al Hazzouri A. Practical guide to honest causal forests for identifying heterogeneous treatment effects. Am J Epidemiol. 2023;192(7):1155–65. source

      Not checked in the free run.

    • cites [38] Cui Y, Kosorok MR, Sverdrup E, Wager S, Zhu R. Estimating heterogeneous treatment effects with right-censored data via causal survival forests. J R Stat Soc Ser B Stat Methodol. 2023;85(2):179–211. source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    ML and AI allow for integration of multiple data types, including EHRs, genomic data, and environmental exposures, and of multimodal data, providing a comprehensive perspective of the determinants of health outcomes across different stages of the life course.

    “Additionally, ML and AI allow for integration of multiple data types, including EHRs, genomic data, and environmental exposures, and of multimodal data, such as kinds of information from individual’s various types of records, thus providing a comprehensive perspective of the determinants of health outcomes across the different stages of life course [32].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [32] Chen RJ, Lu MY, Williamson DFK, Chen TY, Lipkova J, Noor Z, Shaban M, Shady M, Williams M, Joo B, Mahmood F. Pan-cancer integrative histology-genomic analysis via multimodal deep learning. Cancer Cell. 2022;40(8):865–878 e866. source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Advanced ML techniques, such as causal forests, can directly estimate heterogeneous treatment effects and minimize bias in observational studies.

    “Additionally, advanced ML techniques, such as causal forests, can directly estimate heterogeneous treatment effects and minimize bias in observational studies [37–39].”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites [37] Jawadekar N, Kezios K, Odden MC, Stingone JA, Calonico S, Rudolph K, Zeki Al Hazzouri A. Practical guide to honest causal forests for identifying heterogeneous treatment effects. Am J Epidemiol. 2023;192(7):1155–65. source

      Not checked in the free run.

    • cites [38] Cui Y, Kosorok MR, Sverdrup E, Wager S, Zhu R. Estimating heterogeneous treatment effects with right-censored data via causal survival forests. J R Stat Soc Ser B Stat Methodol. 2023;85(2):179–211. source

      Not checked in the free run.

    • cites [39] Credit K, Lehnert M. A structured comparison of causal machine learning methods to assess heterogeneous treatment effects in spatial data. In: Journal of Geographical Systems. 2023.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Transformers, exemplified by the Bidirectional Encoder Representations from Transformers (BERT) model, have exhibited superior performance across a wide range of natural language processing tasks and have been successfully adapted for time series analysis.

    “Transformers, exemplified by the Bidirectional Encoder Representations from Transformers (BERT) model, have exhibited superior performance across a wide range of natural language processing tasks and have been successfully adapted for time series analysis [45, 46].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [45] Homburg M, Meijer E, Berends M, Kupers T, Olde Hartman T, Muris J, de Schepper E, Velek P, Kuiper J, Berger M, Peters L. A natural language processing model for COVID-19 detection based on dutch general practice electronic health records by using bidirectional encoder representations from transformers: development and validation study. J Med Internet Res. 2023;25:e49944. source

      Not checked in the free run.

    • cites [46] Stojanov R, Popovski G, Cenikj G, Korousic Seljak B, Eftimov T. A Fine-tuned bidirectional encoder representations from transformers model for food named-entity recognition: algorithm development and validation. J Med Internet Res. 2021;23(8):e28229. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    ML and AI can incorporate data from various sources, such as EHRs, genomic data, environmental exposures, and structured and unstructured text, audio, image, and video data, to provide a more comprehensive understanding of factors influencing health outcomes.

    “Integrating multiple and multimodal data types: ML and AI can incorporate data from various sources, such as EHRs, genomic data, and environmental exposures, as well as structured and unstructured text, audio, image, and video data, to provide a more comprehensive understanding of the factors influencing health outcomes.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Recurrent neural networks (RNNs) and long short-term memory (LSTM) networks have been extensively employed for temporal data analysis, enabling the capture of dependencies and patterns over time.

    “Recurrent neural networks (RNNs) and long short-term memory (LSTM) networks have been extensively employed for temporal data analysis, enabling the capture of dependencies and patterns over time [42–44].”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [42] Mao S, Sejdic E. A review of recurrent neural network-based methods in computational physiology. IEEE Trans Neural Netw Learn Syst. 2023;34(10):6983–7003. source

      Not checked in the free run.

    • cites [43] Wu Z, Tian Y, Li M, Wang B, Quan Y, Liu J. Prediction of air pollutant concentrations based on the long short-term memory neural network. J Hazard Mater. 2024;465:133099. source

      Not checked in the free run.

    • cites [44] Liu X, Zhang X, Wang R, Liu Y, Hadiatullah H, Xu Y, Wang T, Bendl J, Adam T, Schnelle-Kreis J, Querol X. High-precision microscale particulate matter prediction in diverse environments using a long short-term memory neural network and street view imagery. Environ Sci Technol. 2024;58(8):3869–82. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    When dealing with dynamic variables, it is essential to retrain and reevaluate the models to account for new trends and changes, requiring ongoing monitoring and updates to maintain accuracy and relevance.

    “When dealing with dynamic variables, it is essential to retrain and reevaluate the models to account for new trends and changes, requiring ongoing monitoring and updates to maintain accuracy and relevance.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    Life-course studies often involve data collected using different methods, at different time points, and from diverse populations, and ensuring the comparability and interoperability of these data is crucial for developing robust and generalizable ML and AI models.

    “Life-course studies often involve data collected using different methods, at different time points, and from diverse populations. Ensuring the comparability and interoperability of these data is crucial for developing robust and generalizable ML and AI models.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationnone given

    Life-course studies often involve factors with complex interactions and dynamic nature, as well as social phenomena that are inherently difficult to quantify and model.

    “Furthermore, life-course studies often involve factors with complex interactions and dynamic nature, as well as social phenomena that are inherently difficult to quantify and model.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisecitation

    Integrating ML and AI findings with existing domain knowledge and expert interpretation is essential for ensuring the validity and relevance of the results.

    “Integrating ML and AI findings with existing domain knowledge and expert interpretation is essential for ensuring the validity and relevance of the results [59, 60].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [59] Murdoch WJ, Singh C, Kumbier K, Abbasi-Asl R, Yu B. Definitions, methods, and applications in interpretable machine learning. Proc Natl Acad Sci U S A. 2019;116(44):22071–80. source

      Not checked in the free run.

    • cites [60] Littmann M, Selig K, Cohen-Lavi L, Frank Y, Hönigschmid P, Kataka E, Mösch A, Qian K, Ron A, Schmid S, et al. Validity of machine learning in biology and medicine increased through collaborations across fields of expertise. Nat Mach Intell. 2020;2(1):18–24. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductreasoning

    This perspective paper is not a systematic review and its narrative approach may be subject to selection bias and may not capture all relevant studies or viewpoints.

    “It is important to acknowledge that this perspective paper, while striving for a balanced and comprehensive discussion, is not a systematic review. As such, the information presented may be subject to selection bias. Our narrative approach, while broad, may not capture all relevant studies or viewpoints.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    Multidisciplinary teams of epidemiologists, data scientists, and public health professionals can drive innovation, uncover novel insights, and improve population health outcomes.

    “By working together, these multidisciplinary teams can drive innovation, uncover novel insights, and ultimately improve population health outcomes.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Epidemiologists contribute deep understanding of biological, social, and environmental factors influencing health trajectories; data scientists bring advanced computational and analytical skills; public health professionals provide insights into practical implications and translational potential.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    The integration of ML and AI in life-course epidemiology presents several challenges and ethical considerations that must be addressed to ensure the responsible and effective use of these technologies.

    “The integration of ML and AI in life-course epidemiology presents several challenges and ethical considerations that must be addressed to ensure the responsible and effective use of these technologies.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    Promoting transparency, consistency, and rigor in the use of ML and AI enhances the credibility and impact of life course research.

    “By promoting transparency, consistency, and rigor in the use of ML and AI, the credibility and impact of life course research can be enhanced.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    Collaborations between epidemiologists, data scientists, and public health professionals enable the exchange of knowledge, skills, and expertise necessary to develop and apply cutting-edge ML and AI techniques to complex life-course data.

    “These collaborations will enable the exchange of knowledge, skills, and expertise necessary to develop and apply cutting-edge ML and AI techniques to complex life-course data.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Strategies such as adopting synthetic data for training models provide new opportunities to improve the diversity and robustness of ML and AI models by reducing patient privacy concerns and facilitating data sharing while maintaining the original distribution of data.

    “Strategies such as adopting synthetic data for training models provide new opportunities to improve the diversity and robustness of ML and AI models by reducing patient privacy concerns and facilitating data sharing while maintaining the original distribution of data [63].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [63] Chen RJ, Lu MY, Chen TY, Williamson DFK, Mahmood F. Synthetic data in machine learning for medicine and healthcare. Nat Biomed Eng. 2021;5(6):493–7. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationnone given

    Leveraging ML and AI to analyze diverse datasets can yield novel insights, improve disease risk prediction, and inform the development of targeted interventions.

    “Leveraging these powerful technologies to analyze diverse datasets can yield novel insights, improve disease risk prediction, and inform the development of targeted interventions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    The realization of the potential of ML and AI in life-course epidemiology is contingent upon addressing significant challenges including data quality, model interpretability, bias, privacy, and equity.

    “However, the realization of this potential is contingent upon addressing the significant challenges associated with the use of ML and AI, including issues related to data quality, model interpretability, bias, privacy, and equity.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Unlike certain health specialties such as dermatology or ophthalmology, where ML and AI have been successfully adopted due to their reliance on visual evaluation and pattern recognition, the application of ML and AI in epidemiology presents unique challenges.

    “Unlike certain health specialties such as dermatology or ophthalmology, where ML and AI have been successfully adopted due to their reliance on visual evaluation and pattern recognition [11], the application of ML and AI in epidemiology presents unique challenges.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [11] Lee EY, Maloney NJ, Cheng K, Bach DQ. Machine learning for precision dermatology: advances, opportunities, and outlook. J Am Acad Dermatol. 2021;84(5):1458–9. source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Some data sources used for training ML models, such as EHRs collected for administrative purposes, might not be gathered with the necessary frequency, granularity, or bandwidth that align with information needs of science and learning, and may present challenges in generating accurate and reliable algorithms.

    “Some data sources used for training ML models, such as EHRs collected for administrative purposes, might not be gathered with the necessary frequency, granularity, or bandwidth that align with information needs of science and learning, and may therefore present challenges in generating accurate and reliable algorithms [54].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [54] London AJ. Artificial intelligence in medicine: overcoming or recapitulating structural challenges to improving patient care? Cell Rep Med. 2022;3(5):100622. source

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    ML models have limited prediction value when applied to complex social phenomena, in contrast to their success in physical and biological contexts.

    “This limited prediction value highlights the potential limitations of ML models when applied to complex social phenomena, contrasting with their success in physical and biological contexts [66, 67].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [66] Salganik MJ, Lundberg I, Kindel AT, Ahearn CE, Al-Ghoneim K, Almaatouq A, Altschul DM, Brand JE, Carnegie NB, Compton RJ, et al. Measuring the predictability of life outcomes with a scientific mass collaboration. Proc Natl Acad Sci U S A. 2020;117(15):8398–403. source

      Not checked in the free run.

    • cites [67] Buchholz O, Grote T. Predicting and explaining with machine learning models: social science as a touchstone. Stud Hist Philos Sci. 2023;102:60–9. source

      Not checked in the free run.

  • May combine two claims.

    Interpretationnone given

    While ML and AI can identify novel patterns and associations in data, they do not necessarily provide insights into the underlying biological or clinical mechanisms.

    “While ML and AI can identify novel patterns and associations in data, they do not necessarily provide insights into the underlying biological or clinical mechanisms.”
    Quote in the paperSubstantial claim
49 of 49

  • The paper relies on this without support.

    Applying ML and AI techniques to large-scale epidemiological data can be computationally intensive and require specialized expertise in data science and programming.

    “Applying ML and AI techniques to large-scale epidemiological data can be computationally intensive and require specialized expertise in data science and programming.”

  • The paper relies on this without support.

    Access to high-performance computing resources and qualified personnel may be a barrier for some research groups, particularly in low- and middle-income settings.

    “Access to high-performance computing resources and qualified personnel may be a barrier for some research groups, particularly in low- and middle-income settings.”

  • The paper relies on this without support.

    ML and AI models have demonstrated methodological efficacy.

    “Despite the demonstrated methodological efficacy of ML and AI models”

  • The paper relies on this without support.

    Epidemiologists contribute deep understanding of biological, social, and environmental factors influencing health trajectories; data scientists bring advanced computational and analytical skills; public health professionals provide insights into practical implications and translational potential.

    “Epidemiologists contribute a deep understanding of the biological, social, and environmental factors influencing health trajectories, while data scientists bring advanced computational and analytical skills. Public health professionals provide invaluable insights into the practical implications and translational potential of research findings.”
4 of 4

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Existing AI tools have been utilized to facilitate personalized, self-paced learning, uncover insights for targeted interventions, strengthen decision-making and clinical reasoning in nursing students, provide AI-assisted diagnostics, improve academic writing, and apply advancements in patient care.

    “For instance, existing AI tools have been utilized to facilitate personalized, self-paced learning within courses, uncover new insights for targeted interventions, strengthen decision-making processes and clinical reasoning skills in nursing students, provide AI-assisted diagnostics, improve academic writing, and apply these advancements in patient care [26–31].”

    From Exploring the integration of AI and national quality courses in China: a study on teaching practices in nursing smart education (2025) · cites this paper as [27]

    The Current applications of ML and AI in epidemiology section (page not provided) describes AI-assisted disease prediction and early detection, targeted interventions, and tools used to support clinical care decisions. It does not address personalized, self-paced learning, nursing students’ clinical reasoning, or academic writing; those parts of the sentence are not supported by this source. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

Cites

Flagged first, then by citations

3 of the 72 references couldn’t be matched to a known paper.

Research works printed but not linked: 3

  1. Elder Jr GH, Shanahan MJ. The Life Course and Human Development. In: Damon W, Lerner RM, editors. Handbook of Child Psychology. Volume 1, edn. New Jersey: Wiley; 2007.
  2. Gandhi A, Hosanagar K, Singh A. Machine learning instrument variables for causal inference. In: EC’20: Proceedings of the 21st ACM Conference on Economics and Computation: 2019. 2019.
  3. Hamilton AJ, Strauss AT, Martinez DA, Hinson JS, Levin S, Lin G, Klein EY. Machine learning and artificial intelligence: applications in healthcare epidemiology. Antimicrob Steward Healthc Epidemiol. 2021;1(1):e28. · DOI · Paper page

Cited by

Sorted by citations

Related

Authors
Chen, S., Yu, J., Chamouni, S., Wang, Y., & Li, Y.
Affiliations
  • London School of Hygiene & Tropical Medicine (GB)
  • Chinese University of Hong Kong (HK)
  • University College London (GB)
  • Karolinska Institutet (SE)
Corresponding author
  • Chen, Shanquan
  • Li, Yunfei
ORCID
Venue
BMC Medicine
Publisher
Springer Science and Business Media LLC
Published
2024-09-02
Volume
22
Issue
1
Pages
354-354
ISSN
1741-7015 (ISSN-L)
DOI
10.1186/s12916-024-03566-x
PMC
PMC11367811
OpenAlex
W4402134928
Cited by
73
References
72
Citations by year
2026: 38 · 2025: 32 · 2024: 3
FWCI
2.51
Type
journal article
Language
en
Open access
Licence
Locations
  • BMC Medicine (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • LSHTM Research Online (London School of Hygiene and Tropical Medicine) (acceptedVersion, other-oa)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Funders
Karolinska Institute
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
Keywords
machine learning, artificial intelligence, life course epidemiology, sensitive period, causal inference, disease risk prediction, multimodal data integration, model interpretability, algorithmic bias, data privacy, health equity, public health interventions, lifespan development, interdisciplinary collaboration
MeSH terms
Machine Learning, Artificial Intelligence, Humans, Public Health

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-07.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-07.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review