Explainable AI for Data-Driven Design of High-Dimensional Predictive Studies

Bruce Guthrie; Damian Machlanski; Ewen M Harrison; Junyu Yan; Kurt Butler; Panagiotis Dimitrakopoulos; Sotirios A Tsaftaris

arxiv: 2605.22243 · v1 · pith:DBW57YHWnew · submitted 2026-05-21 · 💻 cs.LG · cs.AI· stat.AP

Explainable AI for Data-Driven Design of High-Dimensional Predictive Studies

Junyu Yan , Damian Machlanski , Kurt Butler , Panagiotis Dimitrakopoulos , Ewen M Harrison , Bruce Guthrie , Sotirios A Tsaftaris This is my paper

Pith reviewed 2026-05-22 08:20 UTC · model grok-4.3

classification 💻 cs.LG cs.AIstat.AP

keywords explainable AICox proportional hazardspredictive modelingfeature interactiondata-driven designfall predictionhealth data analysis

0 comments

The pith

An explainable AI system recommends feature exclusions, non-linear terms, and interactions that raise a Cox model's C-index from 0.805 to 0.815 on 245,614 patients.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces an Exploratory AI Recommender that first trains flexible black-box models to detect complex patterns in high-dimensional health data and then applies explainable AI methods to convert those patterns into concrete design choices for traditional statistical models. The choices take three forms: dropping unhelpful features, adding non-linear transformations for selected variables, and inserting specific pairwise interactions. When these suggestions are added to a baseline Cox Proportional Hazards model for time-to-first-fall prediction, both discrimination and calibration improve on the original cohort and on two additional public datasets. All of the recommended changes receive independent support in the medical literature.

Core claim

The Exploratory AI Recommender captures intricate data relationships with flexible AI models and then uses explainable AI to translate those relationships into three recommendation types—feature exclusion, non-linear terms, and feature interactions—which, when incorporated into a Cox Proportional Hazards model, raise its C-index from 0.805 (95% CI 0.798-0.812) to 0.815 (95% CI 0.809-0.822) while also improving calibration on a cohort of 245,614 patients.

What carries the argument

The Exploratory AI Recommender framework, which combines flexible AI modelling to detect patterns with explainable AI techniques to produce actionable recommendations for feature exclusion, non-linear terms, and interactions.

If this is right

Transparent statistical models can be systematically upgraded with data-driven suggestions while retaining clinical interpretability.
The three recommendation types—exclusions, non-linear terms, and interactions—can be applied across multiple high-dimensional prediction tasks in health data.
All suggestions generated by the method were corroborated by existing literature, indicating the framework can surface known relationships automatically.
The same pipeline succeeded on two additional public datasets, supporting broader use in predictive study design.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Similar recommender pipelines could reduce the manual trial-and-error currently required when building risk models in other high-dimensional domains such as genomics or claims data.
Embedding the recommender inside electronic health record systems might allow clinicians to iterate on model design without deep statistical expertise.
The approach could be extended to generate ranked lists of recommendations so that analysts can choose how many to adopt based on resource constraints.

Load-bearing premise

The patterns found by the flexible AI models reflect stable, clinically meaningful relationships that improve performance on new data instead of capturing noise or dataset-specific artifacts.

What would settle it

Re-training and testing the same augmented Cox model on an independent cohort drawn from a different health-care system yields a C-index no higher than the baseline model's 0.805.

read the original abstract

Predictive modelling is important for health data analysis and data-driven clinical decision-making. However, predictive studies are challenging to design optimally by hand when tens or even hundreds of features require selection, transformation, or interaction modelling. While complex machine learning models offer high performance, their "black-box" nature limits the clinical trust, transparency, and interpretability required for decision-making. We developed and evaluated an Exploratory AI Recommender that provides data-driven recommendations to improve predictive performance of existing interpretable statistical models. The developed framework uses flexible AI modelling to capture complex data patterns and explainable AI techniques to translate the patterns into three recommendation types: feature exclusion, non-linear terms, and feature interactions. We evaluated the framework by comparing predictive performance of a baseline (i.e., no interactions or non-linear terms) Cox Proportional Hazards (CPH) model against an augmented CPH incorporating recommendations suggested by our method. The primary analysis predicts the time to the first occurrence of a fall or related injury in 245,614 patients. Our method recommended excluding 23 features, including non-linear terms for two features, and including 221 suggested feature interactions. The C-index improved from 0.805 (95% CI 0.798-0.812) to 0.815 (95% CI 0.809-0.822), and so did calibration (intercept: -0.006 to 0.003; slope: 1.063 to 0.950). All recommendations were supported by existing literature. The method also proved effective on two additional public datasets, demonstrating wider applicability. The proposed Exploratory AI Recommender demonstrates the potential of explainable AI and data-driven study design to improve the process of developing, and the performance of high-dimensional transparent predictive models.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper gives a workable XAI pipeline for suggesting feature drops, non-linear terms, and interactions to improve Cox models on high-dimensional health data, with a modest C-index gain on a large falls dataset, but the evaluation leaves open whether the recommendations were generated without leaking into the test numbers.

read the letter

Hi, the core of this is a straightforward workflow: train flexible models on clinical data, pull out XAI signals to recommend 23 feature exclusions, two non-linear terms, and 221 interactions, then fold those into a Cox model and measure the lift. On the main 245k-patient falls cohort they report C-index moving from 0.805 to 0.815 with tighter calibration, and the suggestions line up with published clinical knowledge. They repeat the exercise on two public datasets, which is a reasonable check for broader use. That combination of XAI-driven design advice plus concrete numbers on real survival data is the part worth noting; it gives practitioners a template for letting data guide model tweaks while staying inside an interpretable framework. The numbers are small but directionally consistent, and the literature backing adds some face validity. The soft spot is the evaluation pipeline. The abstract does not describe whether the full recommendation step was run inside training folds only or on the entire cohort before final testing. If the latter, the 0.01 C-index improvement and calibration shift could partly reflect overfitting or post-hoc selection rather than stable structure. Without those details it is hard to judge how much of the gain would hold on fresh data. The paper is aimed at clinical statisticians and ML researchers who build or maintain transparent risk models and want a data-driven assist with high-dimensional feature work. A reader already working on survival analysis or XAI applications in medicine would pick up a usable example and some practical numbers. It is solid enough on the empirical side to merit a serious referee, mainly to press for clearer cross-validation of the recommendation stage and any sensitivity checks. I would send it out for review rather than desk reject.

Referee Report

1 major / 0 minor

Summary. The manuscript introduces an Exploratory AI Recommender that employs flexible AI models and explainable AI techniques to derive data-driven recommendations—feature exclusions, non-linear terms, and interactions—for enhancing interpretable statistical models such as Cox Proportional Hazards (CPH). On a primary cohort of 245,614 patients predicting time to first fall or injury, the augmented CPH achieves a C-index increase from 0.805 (95% CI 0.798-0.812) to 0.815 (95% CI 0.809-0.822) with improved calibration; the approach is also tested on two public datasets, and all 23 exclusions, 2 non-linear terms, and 221 interactions are stated to be supported by existing literature.

Significance. If the reported gains reflect generalizable structure rather than dataset-specific artifacts, the framework could meaningfully advance data-driven design of high-dimensional predictive studies by allowing complex patterns discovered via AI to inform transparent, clinically trusted models. The large primary sample size and multi-dataset evaluation are positive features, but the modest 0.01 C-index lift and reliance on post-hoc literature support limit the strength of the significance claim without stronger internal validation.

major comments (1)

The evaluation procedure does not state whether the full pipeline for generating recommendations (23 feature exclusions, 2 non-linear terms, 221 interactions) was performed on the entire 245,614-patient cohort or isolated within training folds. This detail is load-bearing for the central performance claim, because reporting C-index and calibration on the same data used to derive the augmentations risks overfitting or selection bias that could account for the observed improvement and calibration shift from intercept -0.006/slope 1.063 to intercept 0.003/slope 0.950.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We are grateful to the referee for their careful reading and constructive criticism. We respond to the major comment as follows.

read point-by-point responses

Referee: The evaluation procedure does not state whether the full pipeline for generating recommendations (23 feature exclusions, 2 non-linear terms, 221 interactions) was performed on the entire 245,614-patient cohort or isolated within training folds. This detail is load-bearing for the central performance claim, because reporting C-index and calibration on the same data used to derive the augmentations risks overfitting or selection bias that could account for the observed improvement and calibration shift from intercept -0.006/slope 1.063 to intercept 0.003/slope 0.950.

Authors: We thank the referee for identifying this important omission in our description. The full pipeline was indeed performed on the entire 245,614-patient cohort to derive the data-driven recommendations for the high-dimensional predictive study. We acknowledge that this approach carries a risk of selection bias as noted. To strengthen the manuscript, we will add a new subsection in the Methods describing the procedure and include results from an internal validation using 5-fold cross-validation, where the recommendation generation is restricted to training folds. The revised results will report the C-index and calibration on the test folds to confirm that the performance gains (and calibration improvements) are robust and not attributable to overfitting on the full data. revision: yes

Circularity Check

0 steps flagged

No significant circularity; evaluation metrics independent of recommendation generation

full rationale

The paper generates recommendations (feature exclusions, non-linear terms, interactions) via flexible AI and XAI on the 245,614-patient cohort, then reports C-index and calibration for baseline versus augmented CPH models on the same cohort. No quoted step equates a claimed result to its inputs by construction: the C-index values are computed after separate model fitting and are not redefined from the AI-derived recommendations themselves. The modest improvement is presented as an empirical outcome rather than a definitional or fitted-input equivalence, and the derivation chain remains self-contained against external benchmarks such as literature support for recommendations.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard survival analysis assumptions and the premise that XAI can faithfully surface actionable patterns; no new entities or fitted constants are introduced beyond typical ML hyperparameters.

axioms (1)

domain assumption Cox Proportional Hazards model assumptions (proportional hazards, linear effects unless modified) hold sufficiently for the augmented model to be interpretable and valid.
The paper uses CPH as both baseline and target model throughout the evaluation.

pith-pipeline@v0.9.0 · 5883 in / 1441 out tokens · 42646 ms · 2026-05-22T08:20:23.350220+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

68 extracted references · 68 canonical work pages

[1]

Development and Validation of the American Heart Association’s PREVENT Equations

Khan SS, Matsushita K, Sang Y, et al. Development and Validation of the American Heart Association’s PREVENT Equations. Circulation. 2024;149(6):430-449. doi:10.1161/CIRCULATIONAHA.123.067626

work page doi:10.1161/circulationaha.123.067626 2024
[2]

Development and validation of QRISK3 risk prediction algorithms to estimate future risk of cardiovascular disease: prospective cohort study

Hippisley-Cox J, Coupland C, Brindle P. Development and validation of QRISK3 risk prediction algorithms to estimate future risk of cardiovascular disease: prospective cohort study. BMJ. 2017;357. doi:10.1136/BMJ.J2099

work page doi:10.1136/bmj.j2099 2017
[3]

Predictive models for hospital readmission risk: A systematic review of methods

Artetxe A, Beristain A, Graña M. Predictive models for hospital readmission risk: A systematic review of methods. Comput Methods Programs Biomed. 2018;164:49-64. doi:10.1016/j.cmpb.2018.06.006

work page doi:10.1016/j.cmpb.2018.06.006 2018
[4]

Developing prediction models for clinical use using logistic regression: an overview

Shipe ME, Deppen SA, Farjah F, Grogan EL. Developing prediction models for clinical use using logistic regression: an overview. J Thorac Dis. 2019;11(Suppl 4):S574. doi:10.21037/JTD.2019.01.25

work page doi:10.21037/jtd.2019.01.25 2019
[5]

Survival outcome prediction in cervical cancer: Cox models vs deep-learning model

Matsuo K, Purushotham S, Jiang B, et al. Survival outcome prediction in cervical cancer: Cox models vs deep-learning model. Am J Obstet Gynecol. 2019;220(4):381.e1-381.e14. doi:10.1016/J.AJOG.2018.12.030

work page doi:10.1016/j.ajog.2018.12.030 2019
[6]

A guide to deep learning in healthcare

Esteva A, Robicquet A, Ramsundar B, et al. A guide to deep learning in healthcare. Nature Medicine 2019 25:1. 2019;25(1):24-29. doi:10.1038/s41591-018-0316-z

work page doi:10.1038/s41591-018-0316-z 2019
[7]

Deep learning for healthcare: review, opportunities and challenges

Miotto R, Wang F, Wang S, Jiang X, Dudley JT. Deep learning for healthcare: review, opportunities and challenges. Brief Bioinform. 2018;19(6):1236-1246. doi:10.1093/BIB/BBX044

work page doi:10.1093/bib/bbx044 2018
[8]

Black-box artificial intelligence: an epistemological and critical analysis

Carabantes M. Black-box artificial intelligence: an epistemological and critical analysis. AI & SOCIETY 2019 35:2. 2019;35(2):309-317. doi:10.1007/S00146-019-00888-W

work page doi:10.1007/s00146-019-00888-w 2019
[9]

In humble defense of unexplainable black box prediction models in healthcare

van Royen FS, Weerts HJP, de Hond AAH, et al. In humble defense of unexplainable black box prediction models in healthcare. J Clin Epidemiol. 2026;189. doi:10.1016/j.jclinepi.2025.112013

work page doi:10.1016/j.jclinepi.2025.112013 2026
[10]

Explainable AI in healthcare: to explain, to predict, or to describe? Diagn Progn Res

Carriero A, de Hond A, Cappers B, et al. Explainable AI in healthcare: to explain, to predict, or to describe? Diagn Progn Res. 2025;9(1). doi:10.1186/S41512-025-00213-8

work page doi:10.1186/s41512-025-00213-8 2025
[11]

Barriers to and Facilitators of Artificial Intelligence Adoption in Health Care: Scoping Review

Hassan M, Kushniruk A, Borycki E. Barriers to and Facilitators of Artificial Intelligence Adoption in Health Care: Scoping Review. JMIR Hum Factors. 2024;11(1):e48633. doi:10.2196/48633

work page doi:10.2196/48633 2024
[12]

Artificial Intelligence in Medicine: Chances and Challenges for Wide Clinical Adoption

Varghese J. Artificial Intelligence in Medicine: Chances and Challenges for Wide Clinical Adoption. Visc Med. 2020;36(6):443-449. doi:10.1159/000511930

work page doi:10.1159/000511930 2020
[13]

Ethical and regulatory challenges of AI technologies in healthcare: A narrative review

Mennella C, Maniscalco U, De Pietro G, Esposito M. Ethical and regulatory challenges of AI technologies in healthcare: A narrative review. Heliyon. 2024;10(4):e26297. doi:10.1016/j.heliyon.2024.e26297

work page doi:10.1016/j.heliyon.2024.e26297 2024
[14]

Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence

Hassija V, Chamola V, Mahapatra A, et al. Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence. Cognitive Computation 2023 16:1. 2023;16(1):45-74. doi:10.1007/S12559-023-10179-8

work page doi:10.1007/s12559-023-10179-8 2023
[15]

Chapter 53 The shapley value

Winter E. Chapter 53 The shapley value. Handbook of Game Theory with Economic Applications. 2002;3. doi:10.1016/S1574-0005(02)03016-3

work page doi:10.1016/s1574-0005(02)03016-3 2002
[16]

EXplainable Artificial Intelligence (XAI)—From Theory to Methods and Applications

Ortigossa ES, Gonçalves T, Nonato LG. EXplainable Artificial Intelligence (XAI)—From Theory to Methods and Applications. IEEE Access. 2024;12:80799-80846. doi:10.1109/ACCESS.2024.3409843

work page doi:10.1109/access.2024.3409843 2024
[17]

Automated Stroke Prediction Using Machine Learning: An Explainable and Exploratory Study With a Web Application for Early Intervention

Mridha K, Ghimire S, Shin J, Aran A, Uddin MM, Mridha MF. Automated Stroke Prediction Using Machine Learning: An Explainable and Exploratory Study With a Web Application for Early Intervention. IEEE Access. 2023;11. doi:10.1109/ACCESS.2023.3278273

work page doi:10.1109/access.2023.3278273 2023
[18]

Explainable artificial intelligence (XAI) for exploring spatial variability of lung and bronchus cancer (LBC) mortality rates in the contiguous USA

Ahmed ZU, Sun K, Shelly M, Mu L. Explainable artificial intelligence (XAI) for exploring spatial variability of lung and bronchus cancer (LBC) mortality rates in the contiguous USA. Sci Rep. 2021;11(1). doi:10.1038/s41598-021-03198-8

work page doi:10.1038/s41598-021-03198-8 2021
[19]

Toward Explainable Artificial Intelligence for Precision Pathology

Klauschen F, Dippel J, Keyl P, et al. Toward Explainable Artificial Intelligence for Precision Pathology. Annual Review of Pathology: Mechanisms of Disease. 2024;19. doi:10.1146/annurev-pathmechdis-051222-113147

work page doi:10.1146/annurev-pathmechdis-051222-113147 2024
[20]

Explainable AI for Discovering Disease Biomarkers: A Survey

Temkov S. Explainable AI for Discovering Disease Biomarkers: A Survey. In: Gül ÖM, Fiorini P, Kadry SN, eds. 7th EAI International Conference on Robotic Sensor Networks. Springer Nature Switzerland; 2024:185-192

work page 2024
[21]

Platelet Metabolites as Candidate Biomarkers in Sepsis Diagnosis and Management Using the Proposed Explainable Artificial Intelligence Approach

Yagin FH, Aygun U, Algarni A, Colak C, Al-Hashem F, Ardigò LP. Platelet Metabolites as Candidate Biomarkers in Sepsis Diagnosis and Management Using the Proposed Explainable Artificial Intelligence Approach. J Clin Med. 2024;13(17). doi:10.3390/jcm13175002

work page doi:10.3390/jcm13175002 2024
[22]

Personalized Cardiovascular Disease Risk Prediction Using Random Forest: An Optimized Approach

Burugadda VR, Dutt V, Mamta, Vyas N. Personalized Cardiovascular Disease Risk Prediction Using Random Forest: An Optimized Approach. In: Proceedings - 2023 IEEE World Conference on Applied Intelligence and Computing, AIC 2023. 2023. doi:10.1109/AIC57670.2023.10263915

work page doi:10.1109/aic57670.2023.10263915 2023
[23]

Advancing personalized healthcare: leveraging explainable AI for BPPV risk assessment

Khani M, Luo J, Assadi Shalmani M, Taleban A, Adams J, Friedland DR. Advancing personalized healthcare: leveraging explainable AI for BPPV risk assessment. Health Inf Sci Syst. 2024;13(1):1. doi:10.1007/s13755-024-00317-3

work page doi:10.1007/s13755-024-00317-3 2024
[24]

Simpler is better: Lifting interpretability-performance trade- off via automated feature engineering

Gosiewska A, Kozak A, Biecek P. Simpler is better: Lifting interpretability-performance trade- off via automated feature engineering. Decis Support Syst. 2021;150:113556. doi:10.1016/J.DSS.2021.113556

work page doi:10.1016/j.dss.2021.113556 2021
[25]

Combining machine learning with Cox models to identify predictors for incident post-menopausal breast cancer in the UK Biobank

Liu X, Morelli D, Littlejohns TJ, Clifton DA, Clifton L. Combining machine learning with Cox models to identify predictors for incident post-menopausal breast cancer in the UK Biobank. Scientific Reports 2023 13:1. 2023;13(1):9221-. doi:10.1038/s41598-023-36214-0

work page doi:10.1038/s41598-023-36214-0 2023
[26]

Multimorbidity and adverse outcomes following emergency department attendance: population based cohort study

Blayney MC, Reed MJ, Masterson JA, et al. Multimorbidity and adverse outcomes following emergency department attendance: population based cohort study. BMJ Medicine. 2024;3(1):e000731. doi:10.1136/bmjmed-2023-000731

work page doi:10.1136/bmjmed-2023-000731 2024
[27]

TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods

Collins GS, Moons KGM, Dhiman P, et al. TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385. doi:10.1136/BMJ-2023-078378

work page doi:10.1136/bmj-2023-078378 2024
[28]

Use of the electronic Frailty Index to identify vulnerable patients: A pilot study in primary care

Lansbury LN, Roberts HC, Clift E, Herklots A, Robinson N, Sayer AA. Use of the electronic Frailty Index to identify vulnerable patients: A pilot study in primary care. British Journal of General Practice. 2017;67(664). doi:10.3399/bjgp17X693089

work page doi:10.3399/bjgp17x693089 2017
[29]

Random survival forests

Ishwaran H, Kogalur UB, Blackstone EH, Lauer MS. Random survival forests. Annals of Applied Statistics. 2008;2(3). doi:10.1214/08-AOAS169

work page doi:10.1214/08-aoas169 2008
[30]

Randomized 2 x 2 trial evaluating hormonal treatment and the duration of chemotherapy in node-positive breast cancer patients: An update based on 10 years’ follow-up

Sauerbrei W, Bastert G, Bojar H, et al. Randomized 2 x 2 trial evaluating hormonal treatment and the duration of chemotherapy in node-positive breast cancer patients: An update based on 10 years’ follow-up. Journal of Clinical Oncology. 2000;18(1). doi:10.1200/jco.2000.18.1.94

work page doi:10.1200/jco.2000.18.1.94 2000
[31]

Applied Survival Analysis: Regression Modeling of Time to Event Data: Second Edition

Hosmer DW, Lemeshow S, May S. Applied Survival Analysis: Regression Modeling of Time to Event Data: Second Edition. 2011. doi:10.1002/9780470258019

work page doi:10.1002/9780470258019 2011
[32]

Frailty Is a Risk Factor for Falls in the Older Adults: A Systematic Review and Meta-Analysis

Yang ZC, Lin H, Jiang GH, et al. Frailty Is a Risk Factor for Falls in the Older Adults: A Systematic Review and Meta-Analysis. Journal of Nutrition, Health and Aging. Springer-Verlag Italia s.r.l. 2023;27(6):487-495. doi:10.1007/s12603-023-1935-8

work page doi:10.1007/s12603-023-1935-8 2023
[33]

Accessed October 14, 2025

Prescribing in the elderly | Medicines guidance | BNF | NICE. Accessed October 14, 2025. https://bnf.nice.org.uk/medicines-guidance/prescribing-in-the-elderly/

work page 2025
[34]

Multiple adverse outcomes associated with antipsychotic use in people with dementia: population based matched cohort study

Mok PLH, Carr MJ, Guthrie B, et al. Multiple adverse outcomes associated with antipsychotic use in people with dementia: population based matched cohort study. BMJ. Published online

work page
[35]

doi:10.1136/bmj-2023-076268

work page doi:10.1136/bmj-2023-076268 2023
[36]

Therapeutic dilemma’s: antipsychotics use for neuropsychiatric symptoms of dementia, delirium and insomnia and risk of falling in older adults, a clinical review

Korkatti-Puoskari N, Tiihonen M, Caballero-Mora MA, Topinkova E, Szczerbińska K, Hartikainen S. Therapeutic dilemma’s: antipsychotics use for neuropsychiatric symptoms of dementia, delirium and insomnia and risk of falling in older adults, a clinical review. Eur Geriatr Med. Springer Science and Business Media Deutschland GmbH. 2023;14(4):709-720. doi:10....

work page doi:10.1007/s41999-023-00837-3 2023
[37]

Antipsychotic Prescribing and Falls Risk in Nursing Home Residents with Dementia

D’Alton M, Eustace A. Antipsychotic Prescribing and Falls Risk in Nursing Home Residents with Dementia. Age Ageing. 2024;53(Supplement_4):afae178.216. doi:10.1093/ageing/afae178.216

work page doi:10.1093/ageing/afae178.216 2024
[38]

Risk of fall or fracture with concomitant use of prescription opioids and other medications in osteoarthritis patients

Khan NF, Bykov K, Katz JN, Glynn RJ, Vine SM, Kim SC. Risk of fall or fracture with concomitant use of prescription opioids and other medications in osteoarthritis patients. Pharmacoepidemiol Drug Saf. 2024;33(3). doi:10.1002/pds.5773

work page doi:10.1002/pds.5773 2024
[39]

Relative risks of adverse events among older adults receiving opioids versus NSAIDs after hospital discharge: A nationwide cohort study

Herzig SJ, Anderson TS, Jung Y, Ngo L, Kim DH, McCarthy EP. Relative risks of adverse events among older adults receiving opioids versus NSAIDs after hospital discharge: A nationwide cohort study. PLoS Med. 2021;18(9). doi:10.1371/journal.pmed.1003804

work page doi:10.1371/journal.pmed.1003804 2021
[40]

Antihistamine use and the risk of injurious falls or fracture in elderly patients: a systematic review and meta-analysis

Cho H, Myung J, Suh HS, Kang HY. Antihistamine use and the risk of injurious falls or fracture in elderly patients: a systematic review and meta-analysis. Osteoporosis International. Springer London. 2018;29(10):2163-2170. doi:10.1007/s00198-018-4564-z

work page doi:10.1007/s00198-018-4564-z 2018
[41]

Frailty status and increased risk for falls: The role of anticholinergic burden

Naharci MI, Tasci I. Frailty status and increased risk for falls: The role of anticholinergic burden. Arch Gerontol Geriatr. 2020;90. doi:10.1016/j.archger.2020.104136

work page doi:10.1016/j.archger.2020.104136 2020
[42]

Association between usual alcohol consumption and risk of falls in middle-aged and older Chinese adults

Sun Y, Zhang B, Yao Q, et al. Association between usual alcohol consumption and risk of falls in middle-aged and older Chinese adults. BMC Geriatr. 2022;22(1):1-11. doi:10.1186/S12877- 022-03429-1/FIGURES/2

work page doi:10.1186/s12877- 2022
[43]

Drugs with anticholinergic effects and cognitive impairment, falls and all-cause mortality in older adults: A systematic review and meta- analysis

Ruxton K, Woodman RJ, Mangoni AA. Drugs with anticholinergic effects and cognitive impairment, falls and all-cause mortality in older adults: A systematic review and meta- analysis. Br J Clin Pharmacol. 2015;80(2):209-220. doi:10.1111/bcp.12617

work page doi:10.1111/bcp.12617 2015
[44]

Robust variable selection methods with Cox model-a selective practical benchmark study

Zhang Y, Muller S. Robust variable selection methods with Cox model-a selective practical benchmark study. Brief Bioinform. 2024;25(6). doi:10.1093/BIB/BBAE508

work page doi:10.1093/bib/bbae508 2024
[45]

Fall risk stratification in older adults: low and not-at-risk status still associated with falls and injuries

Montero-Odasso M, Pieruccini-Faria F, Son S, et al. Fall risk stratification in older adults: low and not-at-risk status still associated with falls and injuries. Age Ageing. 2025;54(3). doi:10.1093/AGEING/AFAF064

work page doi:10.1093/ageing/afaf064 2025
[46]

Application of machine learning in predicting survival outcomes involving real-world data: a scoping review

Huang Y, Li J, Li M, Aparasu RR. Application of machine learning in predicting survival outcomes involving real-world data: a scoping review. BMC Med Res Methodol. 2023;23(1). doi:10.1186/s12874-023-02078-1

work page doi:10.1186/s12874-023-02078-1 2023
[47]

Machine Learning for Medicine Must Be Interpretable, Shareable, Reproducible and Accountable by Design

Bektaş AB, Gönen M. Machine Learning for Medicine Must Be Interpretable, Shareable, Reproducible and Accountable by Design. Published online August 22, 2025. http://arxiv.org/abs/2508.16097

work page arXiv 2025
[48]

Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead

Rudin C. Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nat Mach Intell. Nature Research. 2019;1(5):206-215. doi:10.1038/s42256-019-0048-x

work page doi:10.1038/s42256-019-0048-x 2019
[49]

Drug-Disease Graph: Predicting Adverse Drug Reaction Signals via Graph Neural Network with Clinical Data

Kwak H, Lee M, Yoon S, Chang J, Park S, Jung K. Drug-Disease Graph: Predicting Adverse Drug Reaction Signals via Graph Neural Network with Clinical Data. In: Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 12085 LNAI. Springer; 2020:633-644. doi:10.1007/978-3-030-47436-2_48

work page doi:10.1007/978-3-030-47436-2_48 2020
[50]

Prediction of drug-drug interaction events using graph neural networks based feature extraction

Al-Rabeah MH, Lakizadeh A. Prediction of drug-drug interaction events using graph neural networks based feature extraction. Sci Rep. 2022;12(1). doi:10.1038/s41598-022-19999-4

work page doi:10.1038/s41598-022-19999-4 2022
[51]

Katzman, Uri Shaham, Alexander Cloninger, Jonathan Bates, Tingting Jiang, and Yuval Kluger

Katzman JL, Shaham U, Cloninger A, Bates J, Jiang T, Kluger Y. DeepSurv: Personalized treatment recommender system using a Cox proportional hazards deep neural network. BMC Med Res Methodol. 2018;18(1). doi:10.1186/s12874-018-0482-1

work page doi:10.1186/s12874-018-0482-1 2018
[52]

Axiomatic attribution for deep networks

Sundararajan M, Taly A, Yan Q. Axiomatic attribution for deep networks. 34th International Conference on Machine Learning, ICML 2017. 2017;7:5109-5118

work page 2017
[53]

SurvSHAP(t): Time-dependent explanations of machine learning survival models

Krzyziński M, Spytek M, Baniecki H, Biecek P. SurvSHAP(t): Time-dependent explanations of machine learning survival models. Knowl Based Syst. 2023;262:110234. doi:10.1016/J.KNOSYS.2022.110234

work page doi:10.1016/j.knosys.2022.110234 2023
[54]

Basic Statistics for the Health Sciences

Kuzma JW. Basic Statistics for the Health Sciences. Vol 4. 1992

work page 1992
[55]

Higher-Order Feature Attribution: Bridging Statistics, Explainable AI, and Topological Signal Processing

Butler K, Feng G, Djurić PM. Higher-Order Feature Attribution: Bridging Statistics, Explainable AI, and Topological Signal Processing. ICASSP 2026 - 2026 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Published online May 3, 2026:5261-5265. doi:10.1109/ICASSP55912.2026.11461829

work page doi:10.1109/icassp55912.2026.11461829 2026
[56]

Regularization and Variable Selection Via the Elastic Net

Zou H, Hastie T. Regularization and Variable Selection Via the Elastic Net. J R Stat Soc Series B Stat Methodol. 2005;67(2):301-320. doi:10.1111/J.1467-9868.2005.00503.X

work page doi:10.1111/j.1467-9868.2005.00503.x 2005
[57]

McCartney G, Hoggett R. How well does the Scottish Index of Multiple Deprivation identify income and employment deprived individuals across the urban-rural spectrum and between local authorities? Public Health. 2023;217:26-32. doi:10.1016/J.PUHE.2023.01.009

work page doi:10.1016/j.puhe.2023.01.009 2023
[58]

Development and validation of an electronic frailty index in a national health maintenance organization

Sikron FH, Schenker R, Koom Y, et al. Development and validation of an electronic frailty index in a national health maintenance organization. Aging. 2024;16(20):13025-13038. doi:10.18632/AGING.206141

work page doi:10.18632/aging.206141 2024
[59]

A chronological map of 308 physical and mental health conditions from 4 million individuals in the English National Health Service

Kuan V, Denaxas S, Gonzalez-Izquierdo A, et al. A chronological map of 308 physical and mental health conditions from 4 million individuals in the English National Health Service. Lancet Digit Health. 2019;1(2):e63-e77. doi:10.1016/S2589-7500(19)30012-3

work page doi:10.1016/s2589-7500(19)30012-3 2019
[60]

Fall-Risk-Increasing Drugs: A Systematic Review and Meta-Analysis: I

de Vries M, Seppala LJ, Daams JG, et al. Fall-Risk-Increasing Drugs: A Systematic Review and Meta-Analysis: I. Cardiovascular Drugs. J Am Med Dir Assoc. 2018;19(4). doi:10.1016/j.jamda.2017.12.013

work page doi:10.1016/j.jamda.2017.12.013 2018
[61]

Fall-Risk-Increasing Drugs: A Systematic Review and Meta-Analysis: II

Seppala LJ, Wermelink AMAT, de Vries M, et al. Fall-Risk-Increasing Drugs: A Systematic Review and Meta-Analysis: II. Psychotropics. J Am Med Dir Assoc. 2018;19(4). doi:10.1016/j.jamda.2017.12.098

work page doi:10.1016/j.jamda.2017.12.098 2018
[62]

Fall-Risk-Increasing Drugs: A Systematic Review and Meta-analysis: III

Seppala LJ, van de Glind EMM, Daams JG, et al. Fall-Risk-Increasing Drugs: A Systematic Review and Meta-analysis: III. Others. J Am Med Dir Assoc. 2018;19(4). doi:10.1016/j.jamda.2017.12.099

work page doi:10.1016/j.jamda.2017.12.099 2018
[63]

The many shapley values for model explanation

Sundararajan M, Najmi A. The many shapley values for model explanation. 37th International Conference on Machine Learning, ICML 2020. 2020;PartF168147-12

work page 2020
[64]

Model-Agnostic Interpretability with Shapley Values

Messalas A, Kanellopoulos Y, Makris C. Model-Agnostic Interpretability with Shapley Values. 10th International Conference on Information, Intelligence, Systems and Applications, IISA

work page
[65]

doi:10.1109/IISA.2019.8900669

Published online 2019. doi:10.1109/IISA.2019.8900669

work page doi:10.1109/iisa.2019.8900669 2019
[66]

A unified approach to interpreting model predictions

Lundberg SM, Lee SI. A unified approach to interpreting model predictions. Adv Neural Inf Process Syst. 2017;2017-December

work page 2017
[67]

A rigorous study of integrated gradients method and extensions to internal neuron attributions

Lundstrom DD, Huang T, Razaviyayn M. A rigorous study of integrated gradients method and extensions to internal neuron attributions. In: International Conference on Machine Learning. 2022:14485-14508

work page 2022
[68]

Why Should I Trust You?

Ribeiro MT, Singh S, Guestrin C. “Why should i trust you?” Explaining the predictions of any classifier. In: Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Vols 13-17-August-2016. 2016. doi:10.1145/2939672.2939778 Acknowledgements We acknowledge the support of the UKRI AI programme, and the Engineering and P...

work page doi:10.1145/2939672.2939778 2016

[1] [1]

Development and Validation of the American Heart Association’s PREVENT Equations

Khan SS, Matsushita K, Sang Y, et al. Development and Validation of the American Heart Association’s PREVENT Equations. Circulation. 2024;149(6):430-449. doi:10.1161/CIRCULATIONAHA.123.067626

work page doi:10.1161/circulationaha.123.067626 2024

[2] [2]

Development and validation of QRISK3 risk prediction algorithms to estimate future risk of cardiovascular disease: prospective cohort study

Hippisley-Cox J, Coupland C, Brindle P. Development and validation of QRISK3 risk prediction algorithms to estimate future risk of cardiovascular disease: prospective cohort study. BMJ. 2017;357. doi:10.1136/BMJ.J2099

work page doi:10.1136/bmj.j2099 2017

[3] [3]

Predictive models for hospital readmission risk: A systematic review of methods

Artetxe A, Beristain A, Graña M. Predictive models for hospital readmission risk: A systematic review of methods. Comput Methods Programs Biomed. 2018;164:49-64. doi:10.1016/j.cmpb.2018.06.006

work page doi:10.1016/j.cmpb.2018.06.006 2018

[4] [4]

Developing prediction models for clinical use using logistic regression: an overview

Shipe ME, Deppen SA, Farjah F, Grogan EL. Developing prediction models for clinical use using logistic regression: an overview. J Thorac Dis. 2019;11(Suppl 4):S574. doi:10.21037/JTD.2019.01.25

work page doi:10.21037/jtd.2019.01.25 2019

[5] [5]

Survival outcome prediction in cervical cancer: Cox models vs deep-learning model

Matsuo K, Purushotham S, Jiang B, et al. Survival outcome prediction in cervical cancer: Cox models vs deep-learning model. Am J Obstet Gynecol. 2019;220(4):381.e1-381.e14. doi:10.1016/J.AJOG.2018.12.030

work page doi:10.1016/j.ajog.2018.12.030 2019

[6] [6]

A guide to deep learning in healthcare

Esteva A, Robicquet A, Ramsundar B, et al. A guide to deep learning in healthcare. Nature Medicine 2019 25:1. 2019;25(1):24-29. doi:10.1038/s41591-018-0316-z

work page doi:10.1038/s41591-018-0316-z 2019

[7] [7]

Deep learning for healthcare: review, opportunities and challenges

Miotto R, Wang F, Wang S, Jiang X, Dudley JT. Deep learning for healthcare: review, opportunities and challenges. Brief Bioinform. 2018;19(6):1236-1246. doi:10.1093/BIB/BBX044

work page doi:10.1093/bib/bbx044 2018

[8] [8]

Black-box artificial intelligence: an epistemological and critical analysis

Carabantes M. Black-box artificial intelligence: an epistemological and critical analysis. AI & SOCIETY 2019 35:2. 2019;35(2):309-317. doi:10.1007/S00146-019-00888-W

work page doi:10.1007/s00146-019-00888-w 2019

[9] [9]

In humble defense of unexplainable black box prediction models in healthcare

van Royen FS, Weerts HJP, de Hond AAH, et al. In humble defense of unexplainable black box prediction models in healthcare. J Clin Epidemiol. 2026;189. doi:10.1016/j.jclinepi.2025.112013

work page doi:10.1016/j.jclinepi.2025.112013 2026

[10] [10]

Explainable AI in healthcare: to explain, to predict, or to describe? Diagn Progn Res

Carriero A, de Hond A, Cappers B, et al. Explainable AI in healthcare: to explain, to predict, or to describe? Diagn Progn Res. 2025;9(1). doi:10.1186/S41512-025-00213-8

work page doi:10.1186/s41512-025-00213-8 2025

[11] [11]

Barriers to and Facilitators of Artificial Intelligence Adoption in Health Care: Scoping Review

Hassan M, Kushniruk A, Borycki E. Barriers to and Facilitators of Artificial Intelligence Adoption in Health Care: Scoping Review. JMIR Hum Factors. 2024;11(1):e48633. doi:10.2196/48633

work page doi:10.2196/48633 2024

[12] [12]

Artificial Intelligence in Medicine: Chances and Challenges for Wide Clinical Adoption

Varghese J. Artificial Intelligence in Medicine: Chances and Challenges for Wide Clinical Adoption. Visc Med. 2020;36(6):443-449. doi:10.1159/000511930

work page doi:10.1159/000511930 2020

[13] [13]

Ethical and regulatory challenges of AI technologies in healthcare: A narrative review

Mennella C, Maniscalco U, De Pietro G, Esposito M. Ethical and regulatory challenges of AI technologies in healthcare: A narrative review. Heliyon. 2024;10(4):e26297. doi:10.1016/j.heliyon.2024.e26297

work page doi:10.1016/j.heliyon.2024.e26297 2024

[14] [14]

Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence

Hassija V, Chamola V, Mahapatra A, et al. Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence. Cognitive Computation 2023 16:1. 2023;16(1):45-74. doi:10.1007/S12559-023-10179-8

work page doi:10.1007/s12559-023-10179-8 2023

[15] [15]

Chapter 53 The shapley value

Winter E. Chapter 53 The shapley value. Handbook of Game Theory with Economic Applications. 2002;3. doi:10.1016/S1574-0005(02)03016-3

work page doi:10.1016/s1574-0005(02)03016-3 2002

[16] [16]

EXplainable Artificial Intelligence (XAI)—From Theory to Methods and Applications

Ortigossa ES, Gonçalves T, Nonato LG. EXplainable Artificial Intelligence (XAI)—From Theory to Methods and Applications. IEEE Access. 2024;12:80799-80846. doi:10.1109/ACCESS.2024.3409843

work page doi:10.1109/access.2024.3409843 2024

[17] [17]

Automated Stroke Prediction Using Machine Learning: An Explainable and Exploratory Study With a Web Application for Early Intervention

Mridha K, Ghimire S, Shin J, Aran A, Uddin MM, Mridha MF. Automated Stroke Prediction Using Machine Learning: An Explainable and Exploratory Study With a Web Application for Early Intervention. IEEE Access. 2023;11. doi:10.1109/ACCESS.2023.3278273

work page doi:10.1109/access.2023.3278273 2023

[18] [18]

Explainable artificial intelligence (XAI) for exploring spatial variability of lung and bronchus cancer (LBC) mortality rates in the contiguous USA

Ahmed ZU, Sun K, Shelly M, Mu L. Explainable artificial intelligence (XAI) for exploring spatial variability of lung and bronchus cancer (LBC) mortality rates in the contiguous USA. Sci Rep. 2021;11(1). doi:10.1038/s41598-021-03198-8

work page doi:10.1038/s41598-021-03198-8 2021

[19] [19]

Toward Explainable Artificial Intelligence for Precision Pathology

Klauschen F, Dippel J, Keyl P, et al. Toward Explainable Artificial Intelligence for Precision Pathology. Annual Review of Pathology: Mechanisms of Disease. 2024;19. doi:10.1146/annurev-pathmechdis-051222-113147

work page doi:10.1146/annurev-pathmechdis-051222-113147 2024

[20] [20]

Explainable AI for Discovering Disease Biomarkers: A Survey

Temkov S. Explainable AI for Discovering Disease Biomarkers: A Survey. In: Gül ÖM, Fiorini P, Kadry SN, eds. 7th EAI International Conference on Robotic Sensor Networks. Springer Nature Switzerland; 2024:185-192

work page 2024

[21] [21]

Platelet Metabolites as Candidate Biomarkers in Sepsis Diagnosis and Management Using the Proposed Explainable Artificial Intelligence Approach

Yagin FH, Aygun U, Algarni A, Colak C, Al-Hashem F, Ardigò LP. Platelet Metabolites as Candidate Biomarkers in Sepsis Diagnosis and Management Using the Proposed Explainable Artificial Intelligence Approach. J Clin Med. 2024;13(17). doi:10.3390/jcm13175002

work page doi:10.3390/jcm13175002 2024

[22] [22]

Personalized Cardiovascular Disease Risk Prediction Using Random Forest: An Optimized Approach

Burugadda VR, Dutt V, Mamta, Vyas N. Personalized Cardiovascular Disease Risk Prediction Using Random Forest: An Optimized Approach. In: Proceedings - 2023 IEEE World Conference on Applied Intelligence and Computing, AIC 2023. 2023. doi:10.1109/AIC57670.2023.10263915

work page doi:10.1109/aic57670.2023.10263915 2023

[23] [23]

Advancing personalized healthcare: leveraging explainable AI for BPPV risk assessment

Khani M, Luo J, Assadi Shalmani M, Taleban A, Adams J, Friedland DR. Advancing personalized healthcare: leveraging explainable AI for BPPV risk assessment. Health Inf Sci Syst. 2024;13(1):1. doi:10.1007/s13755-024-00317-3

work page doi:10.1007/s13755-024-00317-3 2024

[24] [24]

Simpler is better: Lifting interpretability-performance trade- off via automated feature engineering

Gosiewska A, Kozak A, Biecek P. Simpler is better: Lifting interpretability-performance trade- off via automated feature engineering. Decis Support Syst. 2021;150:113556. doi:10.1016/J.DSS.2021.113556

work page doi:10.1016/j.dss.2021.113556 2021

[25] [25]

Combining machine learning with Cox models to identify predictors for incident post-menopausal breast cancer in the UK Biobank

Liu X, Morelli D, Littlejohns TJ, Clifton DA, Clifton L. Combining machine learning with Cox models to identify predictors for incident post-menopausal breast cancer in the UK Biobank. Scientific Reports 2023 13:1. 2023;13(1):9221-. doi:10.1038/s41598-023-36214-0

work page doi:10.1038/s41598-023-36214-0 2023

[26] [26]

Multimorbidity and adverse outcomes following emergency department attendance: population based cohort study

Blayney MC, Reed MJ, Masterson JA, et al. Multimorbidity and adverse outcomes following emergency department attendance: population based cohort study. BMJ Medicine. 2024;3(1):e000731. doi:10.1136/bmjmed-2023-000731

work page doi:10.1136/bmjmed-2023-000731 2024

[27] [27]

TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods

Collins GS, Moons KGM, Dhiman P, et al. TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385. doi:10.1136/BMJ-2023-078378

work page doi:10.1136/bmj-2023-078378 2024

[28] [28]

Use of the electronic Frailty Index to identify vulnerable patients: A pilot study in primary care

Lansbury LN, Roberts HC, Clift E, Herklots A, Robinson N, Sayer AA. Use of the electronic Frailty Index to identify vulnerable patients: A pilot study in primary care. British Journal of General Practice. 2017;67(664). doi:10.3399/bjgp17X693089

work page doi:10.3399/bjgp17x693089 2017

[29] [29]

Random survival forests

Ishwaran H, Kogalur UB, Blackstone EH, Lauer MS. Random survival forests. Annals of Applied Statistics. 2008;2(3). doi:10.1214/08-AOAS169

work page doi:10.1214/08-aoas169 2008

[30] [30]

Randomized 2 x 2 trial evaluating hormonal treatment and the duration of chemotherapy in node-positive breast cancer patients: An update based on 10 years’ follow-up

Sauerbrei W, Bastert G, Bojar H, et al. Randomized 2 x 2 trial evaluating hormonal treatment and the duration of chemotherapy in node-positive breast cancer patients: An update based on 10 years’ follow-up. Journal of Clinical Oncology. 2000;18(1). doi:10.1200/jco.2000.18.1.94

work page doi:10.1200/jco.2000.18.1.94 2000

[31] [31]

Applied Survival Analysis: Regression Modeling of Time to Event Data: Second Edition

Hosmer DW, Lemeshow S, May S. Applied Survival Analysis: Regression Modeling of Time to Event Data: Second Edition. 2011. doi:10.1002/9780470258019

work page doi:10.1002/9780470258019 2011

[32] [32]

Frailty Is a Risk Factor for Falls in the Older Adults: A Systematic Review and Meta-Analysis

Yang ZC, Lin H, Jiang GH, et al. Frailty Is a Risk Factor for Falls in the Older Adults: A Systematic Review and Meta-Analysis. Journal of Nutrition, Health and Aging. Springer-Verlag Italia s.r.l. 2023;27(6):487-495. doi:10.1007/s12603-023-1935-8

work page doi:10.1007/s12603-023-1935-8 2023

[33] [33]

Accessed October 14, 2025

Prescribing in the elderly | Medicines guidance | BNF | NICE. Accessed October 14, 2025. https://bnf.nice.org.uk/medicines-guidance/prescribing-in-the-elderly/

work page 2025

[34] [34]

Multiple adverse outcomes associated with antipsychotic use in people with dementia: population based matched cohort study

Mok PLH, Carr MJ, Guthrie B, et al. Multiple adverse outcomes associated with antipsychotic use in people with dementia: population based matched cohort study. BMJ. Published online

work page

[35] [35]

doi:10.1136/bmj-2023-076268

work page doi:10.1136/bmj-2023-076268 2023

[36] [36]

Therapeutic dilemma’s: antipsychotics use for neuropsychiatric symptoms of dementia, delirium and insomnia and risk of falling in older adults, a clinical review

Korkatti-Puoskari N, Tiihonen M, Caballero-Mora MA, Topinkova E, Szczerbińska K, Hartikainen S. Therapeutic dilemma’s: antipsychotics use for neuropsychiatric symptoms of dementia, delirium and insomnia and risk of falling in older adults, a clinical review. Eur Geriatr Med. Springer Science and Business Media Deutschland GmbH. 2023;14(4):709-720. doi:10....

work page doi:10.1007/s41999-023-00837-3 2023

[37] [37]

Antipsychotic Prescribing and Falls Risk in Nursing Home Residents with Dementia

D’Alton M, Eustace A. Antipsychotic Prescribing and Falls Risk in Nursing Home Residents with Dementia. Age Ageing. 2024;53(Supplement_4):afae178.216. doi:10.1093/ageing/afae178.216

work page doi:10.1093/ageing/afae178.216 2024

[38] [38]

Risk of fall or fracture with concomitant use of prescription opioids and other medications in osteoarthritis patients

Khan NF, Bykov K, Katz JN, Glynn RJ, Vine SM, Kim SC. Risk of fall or fracture with concomitant use of prescription opioids and other medications in osteoarthritis patients. Pharmacoepidemiol Drug Saf. 2024;33(3). doi:10.1002/pds.5773

work page doi:10.1002/pds.5773 2024

[39] [39]

Relative risks of adverse events among older adults receiving opioids versus NSAIDs after hospital discharge: A nationwide cohort study

Herzig SJ, Anderson TS, Jung Y, Ngo L, Kim DH, McCarthy EP. Relative risks of adverse events among older adults receiving opioids versus NSAIDs after hospital discharge: A nationwide cohort study. PLoS Med. 2021;18(9). doi:10.1371/journal.pmed.1003804

work page doi:10.1371/journal.pmed.1003804 2021

[40] [40]

Antihistamine use and the risk of injurious falls or fracture in elderly patients: a systematic review and meta-analysis

Cho H, Myung J, Suh HS, Kang HY. Antihistamine use and the risk of injurious falls or fracture in elderly patients: a systematic review and meta-analysis. Osteoporosis International. Springer London. 2018;29(10):2163-2170. doi:10.1007/s00198-018-4564-z

work page doi:10.1007/s00198-018-4564-z 2018

[41] [41]

Frailty status and increased risk for falls: The role of anticholinergic burden

Naharci MI, Tasci I. Frailty status and increased risk for falls: The role of anticholinergic burden. Arch Gerontol Geriatr. 2020;90. doi:10.1016/j.archger.2020.104136

work page doi:10.1016/j.archger.2020.104136 2020

[42] [42]

Association between usual alcohol consumption and risk of falls in middle-aged and older Chinese adults

Sun Y, Zhang B, Yao Q, et al. Association between usual alcohol consumption and risk of falls in middle-aged and older Chinese adults. BMC Geriatr. 2022;22(1):1-11. doi:10.1186/S12877- 022-03429-1/FIGURES/2

work page doi:10.1186/s12877- 2022

[43] [43]

Drugs with anticholinergic effects and cognitive impairment, falls and all-cause mortality in older adults: A systematic review and meta- analysis

Ruxton K, Woodman RJ, Mangoni AA. Drugs with anticholinergic effects and cognitive impairment, falls and all-cause mortality in older adults: A systematic review and meta- analysis. Br J Clin Pharmacol. 2015;80(2):209-220. doi:10.1111/bcp.12617

work page doi:10.1111/bcp.12617 2015

[44] [44]

Robust variable selection methods with Cox model-a selective practical benchmark study

Zhang Y, Muller S. Robust variable selection methods with Cox model-a selective practical benchmark study. Brief Bioinform. 2024;25(6). doi:10.1093/BIB/BBAE508

work page doi:10.1093/bib/bbae508 2024

[45] [45]

Fall risk stratification in older adults: low and not-at-risk status still associated with falls and injuries

Montero-Odasso M, Pieruccini-Faria F, Son S, et al. Fall risk stratification in older adults: low and not-at-risk status still associated with falls and injuries. Age Ageing. 2025;54(3). doi:10.1093/AGEING/AFAF064

work page doi:10.1093/ageing/afaf064 2025

[46] [46]

Application of machine learning in predicting survival outcomes involving real-world data: a scoping review

Huang Y, Li J, Li M, Aparasu RR. Application of machine learning in predicting survival outcomes involving real-world data: a scoping review. BMC Med Res Methodol. 2023;23(1). doi:10.1186/s12874-023-02078-1

work page doi:10.1186/s12874-023-02078-1 2023

[47] [47]

Machine Learning for Medicine Must Be Interpretable, Shareable, Reproducible and Accountable by Design

Bektaş AB, Gönen M. Machine Learning for Medicine Must Be Interpretable, Shareable, Reproducible and Accountable by Design. Published online August 22, 2025. http://arxiv.org/abs/2508.16097

work page arXiv 2025

[48] [48]

Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead

Rudin C. Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nat Mach Intell. Nature Research. 2019;1(5):206-215. doi:10.1038/s42256-019-0048-x

work page doi:10.1038/s42256-019-0048-x 2019

[49] [49]

Drug-Disease Graph: Predicting Adverse Drug Reaction Signals via Graph Neural Network with Clinical Data

Kwak H, Lee M, Yoon S, Chang J, Park S, Jung K. Drug-Disease Graph: Predicting Adverse Drug Reaction Signals via Graph Neural Network with Clinical Data. In: Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 12085 LNAI. Springer; 2020:633-644. doi:10.1007/978-3-030-47436-2_48

work page doi:10.1007/978-3-030-47436-2_48 2020

[50] [50]

Prediction of drug-drug interaction events using graph neural networks based feature extraction

Al-Rabeah MH, Lakizadeh A. Prediction of drug-drug interaction events using graph neural networks based feature extraction. Sci Rep. 2022;12(1). doi:10.1038/s41598-022-19999-4

work page doi:10.1038/s41598-022-19999-4 2022

[51] [51]

Katzman, Uri Shaham, Alexander Cloninger, Jonathan Bates, Tingting Jiang, and Yuval Kluger

Katzman JL, Shaham U, Cloninger A, Bates J, Jiang T, Kluger Y. DeepSurv: Personalized treatment recommender system using a Cox proportional hazards deep neural network. BMC Med Res Methodol. 2018;18(1). doi:10.1186/s12874-018-0482-1

work page doi:10.1186/s12874-018-0482-1 2018

[52] [52]

Axiomatic attribution for deep networks

Sundararajan M, Taly A, Yan Q. Axiomatic attribution for deep networks. 34th International Conference on Machine Learning, ICML 2017. 2017;7:5109-5118

work page 2017

[53] [53]

SurvSHAP(t): Time-dependent explanations of machine learning survival models

Krzyziński M, Spytek M, Baniecki H, Biecek P. SurvSHAP(t): Time-dependent explanations of machine learning survival models. Knowl Based Syst. 2023;262:110234. doi:10.1016/J.KNOSYS.2022.110234

work page doi:10.1016/j.knosys.2022.110234 2023

[54] [54]

Basic Statistics for the Health Sciences

Kuzma JW. Basic Statistics for the Health Sciences. Vol 4. 1992

work page 1992

[55] [55]

Higher-Order Feature Attribution: Bridging Statistics, Explainable AI, and Topological Signal Processing

Butler K, Feng G, Djurić PM. Higher-Order Feature Attribution: Bridging Statistics, Explainable AI, and Topological Signal Processing. ICASSP 2026 - 2026 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Published online May 3, 2026:5261-5265. doi:10.1109/ICASSP55912.2026.11461829

work page doi:10.1109/icassp55912.2026.11461829 2026

[56] [56]

Regularization and Variable Selection Via the Elastic Net

Zou H, Hastie T. Regularization and Variable Selection Via the Elastic Net. J R Stat Soc Series B Stat Methodol. 2005;67(2):301-320. doi:10.1111/J.1467-9868.2005.00503.X

work page doi:10.1111/j.1467-9868.2005.00503.x 2005

[57] [57]

McCartney G, Hoggett R. How well does the Scottish Index of Multiple Deprivation identify income and employment deprived individuals across the urban-rural spectrum and between local authorities? Public Health. 2023;217:26-32. doi:10.1016/J.PUHE.2023.01.009

work page doi:10.1016/j.puhe.2023.01.009 2023

[58] [58]

Development and validation of an electronic frailty index in a national health maintenance organization

Sikron FH, Schenker R, Koom Y, et al. Development and validation of an electronic frailty index in a national health maintenance organization. Aging. 2024;16(20):13025-13038. doi:10.18632/AGING.206141

work page doi:10.18632/aging.206141 2024

[59] [59]

A chronological map of 308 physical and mental health conditions from 4 million individuals in the English National Health Service

Kuan V, Denaxas S, Gonzalez-Izquierdo A, et al. A chronological map of 308 physical and mental health conditions from 4 million individuals in the English National Health Service. Lancet Digit Health. 2019;1(2):e63-e77. doi:10.1016/S2589-7500(19)30012-3

work page doi:10.1016/s2589-7500(19)30012-3 2019

[60] [60]

Fall-Risk-Increasing Drugs: A Systematic Review and Meta-Analysis: I

de Vries M, Seppala LJ, Daams JG, et al. Fall-Risk-Increasing Drugs: A Systematic Review and Meta-Analysis: I. Cardiovascular Drugs. J Am Med Dir Assoc. 2018;19(4). doi:10.1016/j.jamda.2017.12.013

work page doi:10.1016/j.jamda.2017.12.013 2018

[61] [61]

Fall-Risk-Increasing Drugs: A Systematic Review and Meta-Analysis: II

Seppala LJ, Wermelink AMAT, de Vries M, et al. Fall-Risk-Increasing Drugs: A Systematic Review and Meta-Analysis: II. Psychotropics. J Am Med Dir Assoc. 2018;19(4). doi:10.1016/j.jamda.2017.12.098

work page doi:10.1016/j.jamda.2017.12.098 2018

[62] [62]

Fall-Risk-Increasing Drugs: A Systematic Review and Meta-analysis: III

Seppala LJ, van de Glind EMM, Daams JG, et al. Fall-Risk-Increasing Drugs: A Systematic Review and Meta-analysis: III. Others. J Am Med Dir Assoc. 2018;19(4). doi:10.1016/j.jamda.2017.12.099

work page doi:10.1016/j.jamda.2017.12.099 2018

[63] [63]

The many shapley values for model explanation

Sundararajan M, Najmi A. The many shapley values for model explanation. 37th International Conference on Machine Learning, ICML 2020. 2020;PartF168147-12

work page 2020

[64] [64]

Model-Agnostic Interpretability with Shapley Values

Messalas A, Kanellopoulos Y, Makris C. Model-Agnostic Interpretability with Shapley Values. 10th International Conference on Information, Intelligence, Systems and Applications, IISA

work page

[65] [65]

doi:10.1109/IISA.2019.8900669

Published online 2019. doi:10.1109/IISA.2019.8900669

work page doi:10.1109/iisa.2019.8900669 2019

[66] [66]

A unified approach to interpreting model predictions

Lundberg SM, Lee SI. A unified approach to interpreting model predictions. Adv Neural Inf Process Syst. 2017;2017-December

work page 2017

[67] [67]

A rigorous study of integrated gradients method and extensions to internal neuron attributions

Lundstrom DD, Huang T, Razaviyayn M. A rigorous study of integrated gradients method and extensions to internal neuron attributions. In: International Conference on Machine Learning. 2022:14485-14508

work page 2022

[68] [68]

Why Should I Trust You?

Ribeiro MT, Singh S, Guestrin C. “Why should i trust you?” Explaining the predictions of any classifier. In: Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Vols 13-17-August-2016. 2016. doi:10.1145/2939672.2939778 Acknowledgements We acknowledge the support of the UKRI AI programme, and the Engineering and P...

work page doi:10.1145/2939672.2939778 2016