CoChat
Sign Up
Cited by 47Artificial Intelligence in HealthcareDiabetes, Cardiovascular Risks, and LipoproteinsMachine Learning in Healthcare

A combined strategy of feature selection and machine learning to identify predictors of prediabetes

Open access
De Silva, Kushan, Jönsson, Daniel, Demmer, Ryan T·2020·Journal of the American Medical Informatics Association
DOI 10.1093/jamia/ocz204PMC PMC7647289
Open this paper in CoChat — free account

Abstract

Objective: To identify predictors of prediabetes using feature selection and machine learning on a nationally representative sample of the US population. Materials and Methods: We analyzed n = 6346 men and women enrolled in the National Health and Nutrition Examination Survey 2013–2014. Prediabetes was defined using American Diabetes Association guidelines. The sample was randomly partitioned to training (n = 3174) and internal validation (n = 3172) sets. Feature selection algorithms were run on training data containing 156 preselected exposure variables. Four machine learning algorithms were applied on 46 exposure variables in original and resampled training datasets built using 4 resampling methods. Predictive models were tested on internal validation data (n = 3172) and external validation data (n = 3000) prepared from National Health and Nutrition Examination Survey 2011–2012. Model performance was evaluated using area under the receiver operating characteristic curve (AUROC). Predictors were assessed by odds ratios in logistic models and variable importance in others. The Centers for Disease Control (CDC) prediabetes screening tool was the benchmark to compare model performance. Results: Prediabetes prevalence was 23.43%. The CDC prediabetes screening tool produced 64.40% AUROC. Seven optimal (≥ 70% AUROC) models identified 25 predictors including 4 potentially novel associations; 20 by both logistic and other nonlinear/ensemble models and 5 solely by the latter. All optimal models outperformed the CDC prediabetes screening tool (P < 0.05). Discussion: Combined use of feature selection and machine learning increased predictive performance outperforming the recommended screening tool. A range of predictors of prediabetes was identified. Conclusion: This work demonstrated the value of combining feature selection with machine learning to identify a wide range of predictors that could enhance prediabetes prediction and clinical decision-making.

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

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

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

Integrity signals

Registry

Crossref · 10.1093/jamia/ocz204

Cites

1 of 1 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

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

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

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?

  • Supervised machine learning models are trained on patient data including lab results, demographic information, and lifestyle factors to predict likelihood of diabetes onset, progression, and related complications using feature selection algorithms.

    “These models are trained on vast amounts of patient data, including lab results, demographic information, and lifestyle factors, to predict the likelihood of diabetes onset, progression, and related complications using sophisticated feature selection algorithms and machine learning approaches to identify optimal predictors from large datasets (34).”

    From Generative artificial intelligence in predictive analysis of diabetes and its complications: a narrative review (2025) · cites this paper as [34]

    Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

    Experimental — accuracy not yet measured

  • De Silva et al.'s models achieved at least 70% area under the receiver operating characteristic curve, significantly outperforming the Centers for Disease Control and Prevention prediabetes screening tool.

    “Their models achieved ≥70% area under the receiver operating characteristic curve (AUROC), significantly outperforming the Centers for Disease Control and Prevention (CDC) prediabetes screening tool (34).”

    From Generative artificial intelligence in predictive analysis of diabetes and its complications: a narrative review (2025) · cites this paper as [34]

  • De Silva et al. used NHANES data, feature selection on 156 exposure variables, and four machine learning algorithms to identify 25 predictors of prediabetes.

    “De Silva et al. conducted a comprehensive analysis using the National Health and Nutrition Examination Survey (NHANES) data, employing feature selection algorithms on 156 exposure variables and applying four machine learning algorithms to identify 25 predictors of prediabetes (34).”

    From Generative artificial intelligence in predictive analysis of diabetes and its complications: a narrative review (2025) · cites this paper as [34]

  • NHANES data has enabled the development of sophisticated machine learning algorithms for prediabetes prediction, with studies demonstrating superior performance compared to traditional screening tools through feature selection techniques.

    “NHANES data has enabled the development of sophisticated machine learning algorithms for prediabetes prediction, with studies demonstrating superior performance compared to traditional screening tools through feature selection techniques applied to comprehensive exposure variables (34-36).”

    From Generative artificial intelligence in predictive analysis of diabetes and its complications: a narrative review (2025) · cites this paper as [34]

Self-check

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

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

Cites

Flagged first, then by citations

Research works printed but not linked: 42

  1. [1]1. Huang Y, Cai X, Mai W, et al. Association between prediabetes and risk of cardiovascular disease and all-cause mortality: systematic review and meta-analysis. BMJ 2016; 355: i5953.. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. [2]2. Huang Y, Cai X, Qiu M, et al. Prediabetes and the risk of cancer: a meta-analysis. Diabetologia 2014; 57 (11): 2261–9. [DOI] [PubMed] [Google Scholar]
  3. [3]3. Edwards CM, Cusi K.. Prediabetes: a worldwide epidemic. Endocrinol Metab Clin North Am 2016; 45 (4): 751–64. [DOI] [PubMed] [Google Scholar]
  4. [4]4. Bansal N. Prediabetes diagnosis and treatment: a review. World J Diabetes 2015; 6 (2): 296–303. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. [5]5. Dall TM, Narayan KV, Gillespie KB, et al. Detecting type 2 diabetes and prediabetes among asymptomatic adults in the United States: modeling American Diabetes Association versus US Preventive Services Task Force diabetes screening guidelines. Popul Health Metr 2014; 12 (1): 12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. [6]6. Yudkin JS, Montori VM.. The epidemic of pre-diabetes: the medicine and the politics. BMJ 2014; 349: g4485.. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. [7]7. Yudkin JS. Prediabetes: are there problems with this label? Yes, the label creates further problems! Diabetes Care 2016; 39 (8): 1468–71. [DOI] [PubMed] [Google Scholar]
  8. [9]9. Kanat M, DeFronzo RA, Abdul-Ghani MA.. Treatment of prediabetes. World J Diabetes 2015; 6 (12): 1207.. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. [10]10. König D, Hörmann J, Predel HG, et al. A 12-month lifestyle intervention program improves body composition and reduces the prevalence of prediabetes in obese patients. Obes Facts 2018; 11 (5): 393–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. [11]11. Glechner A, Keuchel L, Affengruber L, et al. Effects of lifestyle changes on adults with prediabetes: a systematic review and meta-analysis. Prim Care Diabetes 2018; 12 (5): 393–408. [DOI] [PubMed] [Google Scholar] 12.Centers for Disease Control and Prevention. National Health and Nutrition Examination Survey. https://www.cdc.gov/nchs/nhanes/index.htm Accessed July 1, 2019.
  11. [13]13. Kim JY, Goran MI, Toledo‐Corral CM, et al. Comparing glycemic indicators of prediabetes: a prospective study of obese Latino youth. Pediatr Diabetes 2015; 16 (8): 640–3. [DOI] [PubMed] [Google Scholar]
  12. [14]14. Kavakiotis I, Tsave O, Salifoglou A, et al. Machine learning and data mining methods in diabetes research. Comput Struct Biotechnol J 2017; 15: 104–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. [15]15. Dorcely B, Katz K, Jagannathan R, et al. Novel biomarkers for prediabetes, diabetes, and associated complications. DMSO 2017; 10: 345–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. [16]16. Heikes KE, Eddy DM, Arondekar B, et al. Diabetes risk calculator: a simple tool for detecting undiagnosed diabetes and pre-diabetes. Diabetes Care 2008; 31 (5): 1040–5. [DOI] [PubMed] [Google Scholar]
  15. [17]17. Xin Z, Yuan J, Hua L, et al. A simple tool detected diabetes and prediabetes in rural Chinese. J Clin Epidemiol 2010; 63 (9): 1030–5. [DOI] [PubMed] [Google Scholar]
  16. [18]18. Murdoch TB, Detsky AS.. The inevitable application of big data to health care. JAMA 2013; 309 (13): 1351–2. [DOI] [PubMed] [Google Scholar]
  17. [19]19. Casanova R, Saldana S, Simpson SL, et al. Prediction of incident diabetes in the Jackson Heart Study using high-dimensional machine learning. PLoS One 2016; 11 (10): e0163942.. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. [20]20. Collins GS, Mallett S, Omar O, et al. Developing risk prediction models for type 2 diabetes: a systematic review of methodology and reporting. BMC Med 2011; 9 (1): 103.. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. [21]21. Barber SR, Davies MJ, Khunti K, et al. Risk assessment tools for detecting those with pre-diabetes: a systematic review. Diabetes Res Clin Pract 2014; 105 (1): 1–3. [DOI] [PubMed] [Google Scholar]
  20. [22]22. Neumann U, Genze N, Heider D.. EFS: an ensemble feature selection tool implemented as R-package and web-application. BioData Min 2017; 10 (1): 21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. [23]23. Rahman MM, Davis DN.. Addressing the class imbalance problem in medical datasets. IJMLC 2013; 3 (2): 224. [Google Scholar]
  22. [24]24. Han L, Luo S, Yu J, et al. Rule extraction from support vector machines using ensemble learning approach: an application for diagnosis of diabetes. IEEE J Biomed Health Inform 2015; 19 (2): 728–34. [DOI] [PubMed] [Google Scholar]
  23. [25]25. Mazurowski MA, Habas PA, Zurada JM, et al. Training neural network classifiers for medical decision making: the effects of imbalanced datasets on classification performance. Neural Netw 2008; 21 (2-3): 427–36. [DOI] [PMC free article] [PubMed] [Google Scholar] 26.Centers for Disease Control and Prevention. CDC Prediabetes Screening Test. https://www.cdc.gov/diabetes/prevention/pdf/prediabetestest.pdf Accessed July 1, 2019. 27.R Core Team. R: A language and environment for statistical computing. https://www.r-project.org/ Accessed July 1, 2019. 28.American Diabetes Association. Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 2014; 37 (Suppl 1): S81–90. [DOI] [PubMed] [Google Scholar]
  24. [29]29. Buuren SV. Groothuis-Oudshoorn K. mice: multivariate imputation by chained equations in R. J Stat Softw 2010: 1–68. [Google Scholar]
  25. [30]30. Zhang Y, Huang J, Wang P.. A prediction model for the peripheral arterial disease using NHANES data. Medicine 2016; 16: 95.. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. [31]31. Kursa MB, Rudnicki WR.. Feature selection with the Boruta package. J Stat Softw 2010; 36 (11): 1–3. [Google Scholar]
  27. [32]32. Romanski P, Kotthoff L.. Fselector: Selecting Attributes. Vienna: R Foundation for Statistical Computing; https://cran.r-project.org/web/packages/FSelector/FSelector.pdf Accessed July 1, 2019. [Google Scholar]
  28. [33]33. Friedman J, Hastie T, Tibshirani R.. Regularization paths for generalized linear models via coordinate descent. J Stat Softw 2010; 33 (1): 1.. [PMC free article] [PubMed] [Google Scholar]
  29. [35]35. Alghamdi M, Al-Mallah M, Keteyian S, et al. Predicting diabetes mellitus using SMOTE and ensemble machine learning approach: The Henry Ford ExercIse Testing (FIT) project. PLoS One 2017; 12 (7): e0179805.. [DOI] [PMC free article] [PubMed] [Google Scholar] 36.Lunardon N, Menardi G, Torelli N. ROSE: a package for binary imbalanced learning. R J 2014;6 (1): 79–89. [Google Scholar]
  30. [37]37. Chawla NV, Bowyer KW, Hall LO, et al. SMOTE: synthetic minority over-sampling technique. J Artif Intell Res 2002; 16: 321–57. [Google Scholar]
  31. [38]38. Bekkar M, Djemaa HK, Alitouche TA.. Evaluation measures for models assessment over imbalanced datasets. J Inf Eng Appl 2013; 3 (10). [Google Scholar]
  32. [39]39. Chawla NV. Data mining for imbalanced datasets: an overview In: Data Mining and Knowledge Discovery Handbook. Boston: Springer; 2009: 875–86. [Google Scholar]
  33. [40]40. Jayanthi N, Babu BV, Rao NS.. Survey on clinical prediction models for diabetes prediction. J Big Data 2017; 4 (1): 26. [Google Scholar]
  34. [41]41. Poltavskiy E, Kim DJ, Bang H.. Comparison of screening scores for diabetes and prediabetes. Diabetes Res Clin Pract 2016; 118: 146–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. [42]42. Hanley JA, McNeil BJ.. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982; 143 (1): 29–36. [DOI] [PubMed] [Google Scholar]
  36. [43]43. Suvitaival T, Bondia-Pons I, Yetukuri L, et al. Lipidome as a predictive tool in progression to type 2 diabetes in Finnish men. Metab Clin Exp 2018; 78: 1–2. [DOI] [PubMed] [Google Scholar]
  37. [44]44. Morteza A, Nakhjavani M, Asgarani F, et al. Inconsistency in albuminuria predictors in type 2 diabetes: a comparison between neural network and conditional logistic regression. Transl Res 2013; 161 (5): 397–405. [DOI] [PubMed] [Google Scholar]
  38. [45]45. Choi SB, Kim WJ, Yoo TK, et al. Screening for prediabetes using machine learning models. Comput Math Methods Med 2014; 2014: 1. [DOI] [PMC free article] [PubMed] [Google Scholar] 46.World Health Organization. Definition and Diagnosis of Diabetes Mellitus and Intermediate Hyperglycemia: Report of a WHO/IDF Consultation. Geneva: WHO; 2006. [Google Scholar]
  39. [47]47. Cafri G, Bailey BA.. Understanding variable effects from black box prediction: Quantifying effects in tree ensembles using partial dependence. J Data Sci 2016; 14 (1): 67–95. [Google Scholar]
  40. [48]48. Song X, Qiu M, Wang H, et al. Gender-related affecting factors of prediabetes on its 10-year outcome. BMJ Open Diabetes Res Care 2016; 4 (1): e000169.. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. [49]49. Anderson JP, Parikh JR, Shenfeld DK, et al. Reverse engineering and evaluation of prediction models for progression to type 2 diabetes: an application of machine learning using electronic health records. J Diabetes Sci Technol 2016; 10 (1): 6–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. [50]50. Rathmann W, Kowall B, Heier M, et al. Prediction models for incident type 2 diabetes mellitus in the older population: KORA S4/F4 cohort study. Diabet Med 2010; 27 (10): 1116–23. [DOI] [PubMed] [Google Scholar] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article.

Cited by

Sorted by citations

Authors
De Silva, K., Jönsson, D., & Demmer, R. T.
Affiliations
  • Department of Clinical Sciences, Faculty of Medicine, Lund University, Lund,Sweden
  • Malmö University (SE)
  • University of Minnesota (US)
ORCID
Venue
Journal of the American Medical Informatics Association
Publisher
Oxford University Press (OUP)
Published
2019-12-30
Volume
27
Issue
3
Pages
396-406
ISSN
  • 1067-5027 (ISSN-L)
  • 1527-974X
DOI
10.1093/jamia/ocz204
PMC
PMC7647289
OpenAlex
W2997395189
Cited by
47
References
1
Citations by year
2026: 4 · 2025: 4 · 2024: 7 · 2023: 6 · 2022: 13 · 2021: 7 · 2020: 6
FWCI
7.09
Type
journal article
Language
en
Open access
Licence
academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model (vor)
Locations
  • Journal of the American Medical Informatics Association (publishedVersion)
  • PubMed Central (publishedVersion)
  • PubMed (publishedVersion)
  • Lund University Publications (Lund University) (submittedVersion)
Funders
Swedish Institute scholarship granted to Kushan De Silva for studies at Lund University
Topics
  • Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
  • Health Sciences › Medicine › Endocrinology, Diabetes and Metabolism › Diabetes, Cardiovascular Risks, and Lipoproteins
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
Keywords
prediabetes, risk factors, feature selection, machine learning, NHANES, diabetes screening, model validation, feature importance
MeSH terms
Machine Learning, Adult, Algorithms, Female, Humans, Male, Middle Aged, Models, Theoretical, Nutrition Surveys, Prediabetic State, Risk Factors, ROC Curve, Socioeconomic Factors, Prevalence, Area Under Curve

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