Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 63 of the 64 references it cites were found in a registry or an index, and 1 are not checked yet.

Exploring the potential and limitations of deep learning and explainable AI for longitudinal life course analysis

Open access

Published version of a preprint

Abstract

Background: Understanding the complex interplay between life course exposures, such as adverse childhood experiences and environmental factors, and disease risk is essential for developing effective public health interventions. Traditional epidemiological methods, such as regression models and risk scoring, are limited in their ability to capture the non-linear and temporally dynamic nature of these relationships. Deep learning (DL) and explainable artificial intelligence (XAI) are increasingly applied within healthcare settings to identify influential risk factors and enable personalised interventions. However, significant gaps remain in understanding their utility and limitations, especially for sparse longitudinal life course data and how the influential patterns identified using explainability are linked to underlying causal mechanisms. Methods: We conducted a controlled simulation study to assess the performance of various state-of-the-art DL architectures including CNNs and (attention-based) RNNs against XGBoost and logistic regression. Input data was simulated to reflect a generic and generalisable scenario with different rules used to generate multiple realistic outcomes based upon epidemiological concepts. Multiple metrics were used to assess model performance in the presence of class imbalance and SHAP values were calculated. Results: We find that DL methods can accurately detect dynamic relationships that baseline linear models and tree-based methods cannot. However, there is no one model that consistently outperforms the others across all scenarios. We further identify the superior performance of DL models in handling sparse feature availability over time compared to traditional machine learning approaches. Additionally, we examine the interpretability provided by SHAP values, demonstrating that these explanations often misalign with causal relationships, despite excellent predictive and calibrative performance. Conclusions: These insights provide a foundation for future research applying DL and XAI to life course data, highlighting the challenges associated with sparse healthcare data, and the critical need for advancing interpretability frameworks in personalised public health.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1186/s12889-025-22705-4

Cites

63 of 64 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Predictive performance of models for each of the four LCPs (Period, Repeats, Order and Timing). Model performance, measured by AUPRC scores, shown for each of the three patterns that fall within each LCP. All the AUPRC scores that resulted in each of the three different seeds used to control model training and hyperparameter optimisation are shown. Boxes show the interquartile range, midpoints depict the median values and whiskers show the minimum and maximum values excluding outliers. LR = logistic regression, XGB = XGBoost, MLF = MLSTM-FCN, IT = InceptionTime, LSTMA = LSTMAttention
Fig. 1 Predictive performance of models for each of the four LCPs (Period, Repeats, Order and Timing). Model performance, measured by AUPRC scores, shown for each of the three patterns that fall within each LCP. All the AUPRC scores that resulted in each of the three different seeds used to control model training and hyperparameter optimisation are shown. Boxes show the interquartile range, midpoints depict the median values and whiskers show the minimum and maximum values excluding outliers. LR = logistic regression, XGB = XGBoost, MLF = MLSTM-FCN, IT = InceptionTime, LSTMA = LSTMAttentionCC BY · PubMed Central
Fig. 2 SHAP Beeswarm Plots for LCP Order1. The top panel illustrates SHAP values marginalised over time for each feature, grouped by model, with features ordered by mean absolute SHAP values. The bottom panel visualises SHAP values marginalised over features for each age, also grouped by model. Feature values are colour-coded, ranging from low (“Low”) to high (“High”). LR = logistic regression, XGB = XGBoost, MLF = MLSTM-FCN, IT = InceptionTime, LSTMA = LSTMAttention
Fig. 2 SHAP Beeswarm Plots for LCP Order1. The top panel illustrates SHAP values marginalised over time for each feature, grouped by model, with features ordered by mean absolute SHAP values. The bottom panel visualises SHAP values marginalised over features for each age, also grouped by model. Feature values are colour-coded, ranging from low (“Low”) to high (“High”). LR = logistic regression, XGB = XGBoost, MLF = MLSTM-FCN, IT = InceptionTime, LSTMA = LSTMAttentionCC BY · PubMed Central
Fig. 3 SHAP Comparison for Individuals for LCP Order1. This figure compares SHAP values for two individuals, one positive for the outcome of interest and one negative. The top row includes individual feature heatmaps and SHAP values marginalised over features. The bottom two rows depict the individual-level SHAP values marginalised over time for each individual split by model. SHAP values are shown on a consistent scale across all plots, with colour gradients indicating magnitude and direction (blue to red). LR = logistic regression, XGB = XGBoost, MLF = MLSTM-FCN, IT = InceptionTime, LSTMA = LSTMAttention
Fig. 3 SHAP Comparison for Individuals for LCP Order1. This figure compares SHAP values for two individuals, one positive for the outcome of interest and one negative. The top row includes individual feature heatmaps and SHAP values marginalised over features. The bottom two rows depict the individual-level SHAP values marginalised over time for each individual split by model. SHAP values are shown on a consistent scale across all plots, with colour gradients indicating magnitude and direction (blue to red). LR = logistic regression, XGB = XGBoost, MLF = MLSTM-FCN, IT = InceptionTime, LSTMA = LSTMAttentionCC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

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?

1 more claim about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 61 references

Research works printed but not linked: 7

  1. Zerveas G, Jayaraman S, Patel D, Bhamidipaty A, Eickhoff C. A Transformer-based Framework for Multivariate Time Series Representation Learning. Proc ACM SIGKDD Int Conf Knowl Discov Data Min. 2021;11(21):2114–24.
  2. Qi Q, Luo Y, Xu Z, Ji S, Yang T. Stochastic Optimization of Areas Under Precision-Recall Curves with Provable Convergence. Adv Neural Inf Process Syst. 2021;3:1752–65.
  3. Ismail AA, Gunady MK, Bravo HC, Feizi S. Benchmarking Deep Learning Interpretability in Time Series Predictions. Adv Neural Inf Process Sys. 2020;33:6441–52.
  4. Shrikumar A, Greenside P, Kundaje A. Learning Important Features Through Propagating Activation Differences. In: International conference on machine learning, PMIR. 2017;3145–53
  5. Adebayo J, Gilmer J, Muelly M, Goodfellow IJ, Hardt M, Kim B. Sanity Checks for Saliency Maps. Adv Neural Inf Process Systems. 2018;31.
  6. Muschalik M, Baniecki H, Fumagalli F, Kolpaczki P, Hammer B, Hüllermeier E. shapiq: Shapley Interactions for Machine Learning. In: The Thirty-eight Conference on Neural Information Processing Systems Datasets and Benchmarks Track. Vancouver; 2024.
  7. Heskes T, Sijben E, Bucur IG, Claassen T. Causal Shapley Values: Exploiting Causal Knowledge to Explain Individual Predictions of Complex Models. Adv Neural Inf Process Sys. 2020;33:4778–89.

Cited by

See all 34

Metadata

Authors
Coupland, H., Scheidwasser, N., Katsiferis, A., Davies, M., Flaxman, S., Hulvej Rod, N., Mishra, S., Bhatt, S., & Unwin, H. J. T.
Affiliations
  • University of Copenhagen (DK)
  • University of Oxford (GB)
  • National University of Singapore (SG)
  • University of Bristol (GB)
Corresponding author
Bhatt, Samir
Venue
BMC Public Health
Published
2025-04-24
Volume
25
Issue
1
Pages
1520-1520
ISSN
1471-2458 (ISSN-L)
OpenAlex
W4409733048
Cited by
26
References
61
Citations by year
2026: 17 · 2025: 9
FWCI
10.43
Type
journal article
Language
en
Open access
Licence
Locations
  • BMC Public Health (publishedVersion, cc-by)
  • Research Square (acceptedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Research at the University of Copenhagen (University of Copenhagen) (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Bristol Research (University of Bristol) (submittedVersion, cc-by)
Funders
  • Medical Research Council (MR/X020258/1)
  • Novo Nordisk Fonden (NNF20OC0059309)
  • Danish National Research Foundation (DNRF160)
  • Independent Research Fund Project (1030-00171B)
  • EPSRC (EP/V002910/2)
  • Danish Research Council (reference1030-00171B)
  • European Research Council (no. 101124807)
  • TrygFonden
  • National Research Foundation Singapore (NRF-NRFF15-2023-0010)
  • National Institute for Health and Care Research (NIHR200908)
  • Schmidt Family Foundation (G-22-63345)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
  • Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
  • Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
Keywords
deep learning, explainable AI, life course epidemiology, longitudinal data, data sparsity, SHAP, causal inference, adverse childhood experiences, class imbalance, convolutional neural networks, recurrent neural networks, XGBoost, disease risk prediction
MeSH terms
Deep Learning, Artificial Intelligence, Humans, Longitudinal Studies

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.

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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Exploring the potential and limitations of deep learning and explainable AI for longitudinal life course analysis. https://app.cochat.ai/papers/10.1186/s12889-025-22705-4
The paper
Coupland, H., Scheidwasser, N., Katsiferis, A., Davies, M., Flaxman, S., Hulvej Rod, N., Mishra, S., Bhatt, S., & Unwin, H. J. T. (2025). Exploring the potential and limitations of deep learning and explainable AI for longitudinal life course analysis. BMC Public Health. https://doi.org/10.1186/s12889-025-22705-4

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.