Your privacy, your choice

We use essential cookies to make sure the site can function. We also use optional cookies for advertising, personalisation of content, usage analysis, and social media.

By accepting optional cookies, you consent to the processing of your personal data - including transfers to third parties. Some third parties are outside of the European Economic Area, with varying standards of data protection.

See our privacy policy for more information on the use of your personal data.

for further information and to change your choices.
Skip to main content
Download PDF

Predictive value of machine learning for the progression of gestational diabetes mellitus to type 2 diabetes: a systematic review and meta-analysis

BMC Medical Informatics and Decision Making volume 25, Article number: 18 (2025) Cite this article

Abstract

Background

This systematic review aims to explore the early predictive value of machine learning (ML) models for the progression of gestational diabetes mellitus (GDM) to type 2 diabetes mellitus (T2DM).

Methods

A comprehensive and systematic search was conducted in Pubmed, Cochrane, Embase, and Web of Science up to July 02, 2024. The quality of the studies included was assessed. The risk of bias was assessed through the prediction model risk of bias assessment tool and a graph was drawn accordingly. The meta-analysis was performed using Stata15.0.

Results

A total of 13 studies were included in the present review, involving 11,320 GDM patients and 22 ML models. The meta-analysis for ML models showed a pooled C-statistic of 0.82 (95% CI: 0.79 ~ 0.86), a pooled sensitivity of 0.76 (0.72 ~ 0.80), and a pooled specificity of 0.57 (0.50 ~ 0.65).

Conclusion

ML has favorable diagnostic accuracy for the progression of GDM to T2DM. This provides evidence for the development of predictive tools with broader applicability.

Peer Review reports

Background

Gestational diabetes mellitus (GDM) is characterized by impaired glucose tolerance during pregnancy [1]. The incidence of GDM has increased to 20% due to improved quality of life [2]. It can induce various complications, which bring a heavy burden to GDM patients and the medical system [3]. Compared with their peers, women with a history of GDM have a higher risk of developing type 2 diabetes mellitus (T2DM) [4] and a 7-fold increased risk of T2DM later in life [5], despite having similar baseline levels of impaired glucose tolerance [6]. The incidence of T2DM can be as high as 30–50% within 5–10 years after pregnancy [7, 8]. T2DM may lead to various complications, such as retinopathy, nephropathy, neuropathy, and vascular diseases [9, 10], and patients with a history of GDM are more likely to develop health-threatening diseases in subsequent disability-adjusted life years. The risk of T2DM can be reduced through early identification and healthy living [11,12,13]. A balanced diet that includes meals and snacks, fiber-rich plant foods (legumes, nuts, seeds, herbs), non-starchy vegetables, nutrient-rich carbohydrate foods (whole grains), fish, and unsaturated vegetable oils can greatly benefit a healthy life. Additionally, aerobic exercise and resistance training both improve skeletal muscle, adipose tissue, liver, and pancreatic function; self-monitoring of blood glucose and adherence to medication [14,15,16] are essential for patients with a history of GDM. In recent years, machine learning (ML) has been gradually applied to clinical practice due to its desirable prediction and stratification performance [17, 18], such as the diagnosis of disease status (e.g., tumor lymph node metastasis status [19, 20], and molecular subtype [21]), prognostic prediction (e.g., early tumor recurrence [22]), and prediction of treatment responses [23, 24]. The modeling techniques and variables of ML are diverse, which is one of the reasons why its predictive performance is highly controversial. Hence, this meta-analysis was conducted to explore the accuracy of ML in predicting T2DM risk in patients with a history of GDM and provide references for the development and update of simple risk stratification tools in this field.

Materials and methods

This paper followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statements [25] and was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (registration number: CRD42022333398).

Search strategy

Pubmed, Cochrane, Embase, and Web of Science were comprehensively searched for relevant studies until May 25, 2023, without restrictions on languages and regions. The combination of subject headings and free words was used as the strategy. An additional search of the databases was conducted on July 02, 2024 to minimize the risk of missing new publications. The specific search process is shown in the supplementary material 1.

Inclusion criteria

The inclusion criteria were as follows: (i) study subjects were women with GDM; (ii) studies explicitly described whether GDM occurred during the follow-up period; (iii) in studies of ML, any of the following metrics (Roc, C-statistic, sensitivity/recall, specificity, accuracy, precision, confusion matrix, diagnostic four-compartment table, F1 score) were used to assess the accuracy of ML models. Therefore, at least one of these metrics was included for evaluating the accuracy of ML models in predicting T2DM in GDM patients.

Exclusion criteria

Meanwhile, the studies were excluded for the following reasons: (i) systematic reviews or meta-analyses, consensus statements, guidelines, review articles, and conference abstracts published without peer review; (ii) only risk factors were analyzed without building a complete prediction model; (iii) non-English articles.

Literature screening and data extraction

All retrieved literature was imported into EndNote X9. After duplicates were removed, the titles and abstracts of the remaining studies were preliminarily reviewed to exclude ineligible literature before the full texts were downloaded. After a full-text review, 13 studies were finally included.

A standard spreadsheet was prepared for information extraction, including first author, publication year, country, study type, sample source, sample size of T2DM in the training set, sample size in the training set, external validation, sample size of T2DM in the validation set, sample size in the validation set, method for variable selection, model type, number of model variables, model variables, outcomes, and follow-up duration.

Literature screening and information extraction were performed independently by two researchers (Zhao M and Wei LL) and the results were subsequently cross-checked. Any disputes were tackled through discussion with a third researcher (Ma LD).

Quality assessment

The risk of bias (ROB) of ML models in eligible studies was evaluated using the Prediction Model Risk of Bias Assessment Tool (PROBAST) in four major domains: participants, predictors, outcomes, and statistical analysis. The four domains included two, three, six, and nine questions, respectively. These questions were answered as yes/probable yes (Y/PY), no/probably no (N/PN), or no information (NI). If a domain was answered with at least one N/PN, it was rated as having high risk. In contrast, a domain with all its questions answered as Y/PY was regarded as having low risk. When all domains were at low risk, the overall ROB was rated as low. However, the overall ROB was high when at least one domain was considered at high risk.

Statistical analysis

Data synthesis

A meta-analysis of the metrics (C-statistics and accuracy) was performed using R4.2.0 (R development Core Team, Vienna, http://www.R-project.org). If C-statistics lacked 95% confidence interval (CI) and standard error (SE), the SE was estimated according to the method by Debray TP et al. [26]. For inaccurate original data, we conducted calculations based on sensitivity and specificity as well as the sample size of each molecular subtype and the sample size of modeling. Given the difference in variables and parameters among ML models, the random-effects model was utilized for meta-analysis.

Results

Literature screening

A comprehensive and systematic search was first conducted in Pubmed, Cochrane, Embase, and Web of Science until May 25, 2023. A total of 2, 820 articles were retrieved. After removing 491 duplicates, we reviewed the titles and abstracts of the remaining 2, 329 papers. The full texts of 16 potentially eligible studies were downloaded. After a full-text review, 1 unpublished conference paper was excluded, 1 paper about prediabetes, not GDM was excluded; and 2 papers were excluded for lacking outcome indicators to measure model accuracy. Therefore, 12 original studies were included in the first stage.

In addition, we conducted a supplemental search on July 2, 2024, with 443 publications retrieved from various databases, and 1 newly published study was added after the same screening process. Eventually, a total of 13 studies [27,28,29,30,31,32,33,34,35,36,37,38,39] were included in the present research (Fig. 1).

Fig. 1
figure 1

Flow diagram of study selection

Full size image

Characteristics of included studies

A total of 13 eligible studies were included in the present study, involving 11,320 patients with a history of GDM. During the follow-up period, 1,246 cases progressed to T2DM (Tables 1 and 2). The specific characteristics of the 13 eligible studies are shown in Fig. 1. These studies were published from 2011 to 2024 in eight countries, including Canada, China, Germany, the United States, the United Kingdom, Australia, Italy, and Israel. There were 6 prospective studies [27, 28, 30, 32, 35, 39], 1 randomized controlled trial [34], and 6 retrospective studies [29, 31, 33, 36,37,38]. Among these studies, two studies [28, 35] conducted external validation, and five studies performed cross-validation [27, 31, 33, 34, 39]. The mean follow-up duration was 5 years.

Table 1 Basic characteristics of the included studies
Full size table
Table 2 Modeling information of the included studies
Full size table

Quality assessment

The ROB of the studies included was assessed using PROBAST. The model types included Logistic Regression (LR), Decision Trees (DT), Naive Bayes (NB), Cox, and Random Forest (RF), among which Cox and DT were the predominant types. Of the studies we included, 7 models came from case-control studies, and 2 models came from studies that did not specify the type, so the risk of bias was unclear. 7 models from case-control studies had a high risk of bias in predictor evaluation, and 1 model contained a large number of missing values, and therefore it had a high risk of bias in predictors. The number of missing values could not be determined for 1 model; therefore, the validity of its predictors was unclear. 6 models had an unclear risk of bias in outcomes since it was unable to determine whether the information about the predictors was clear when assessing the outcome. In addition, 17 models were not satisfied with an EPV > 20 or did not have an independent validation set, and therefore they had a high risk of bias in statistical analysis (Fig. 2). Model variables encompassed biomarkers of metabolic disease, lipolytic metabolites, and independent predictors of T2DM.

Fig. 2
figure 2

Quality assessment results of ML models in included studies

Full size image

Results of meta-analysis

C-statistic

23 ML models reported the C-statistic and a random-effects model was used for meta-analysis. The pooled C-statistic was 0.82 (95% CI: 0.79 ~ 0.86) (Fig. 3). A subgroup analysis by model type was performed, which yielded a pooled C-statistic of 0.81 (95% CI: 0.75 ~ 0.88) for LR, 0.85 (95% CI: 0.80 ~ 0.90) for DT, 0.79 (95% CI: 0.71 ~ 0.86) for NB, 0.80 (95% CI: 0.74 ~ 0.85) for Cox, and 0.88 (95% CI: 0.81 ~ 0.94) for RF.

Fig. 3
figure 3

Forest plot of C-statistic

Full size image

Sensitivity and specificity

Due to a significant imbalance in the number of patients with and without T2DM in original studies, the C-statistics of ML models could not adequately explain the accuracy in predicting T2DM. Therefore, meta-analyses of sensitivity and specificity were conducted for the 23 ML models (Fig. 4), which yielded a pooled sensitivity of 0.76 (0.72–0.80) and a pooled specificity of 0.57 (0.50–0.65).

Fig. 4
figure 4

Forest plot of sensitivity and specificity

Full size image

Discussion

This paper summarized relevant studies published in recent years. The overall C-statistic was 0.82 (95%CI: 0.79 ~ 0.86), indicating the favorable performance of ML models in the prediction of the risk of progressing to T2DM in women with a history of GDM, because a C-statistic > 0.7 generally demonstrated an ideal result [40]. High heterogeneity was observed across eligible studies due to the diversity of ML models, mainly in ML algorithms, modeling indicators, risk indicators, and modeling parameters.

The 75-g oral glucose tolerance test (OGTT) and fasting plasma glucose (FPG) tests were recommended by the American Diabetes Association (ADA) for screening T2DM at the 6 to 12 weeks postpartum and every 1–3 years thereafter [41]. Another study also recommended that women at high risk of T2DM should be evaluated for A1C levels [42]. Despite these recommendations, screening of GDM after pregnancy remains notably inadequate, with very low compliance rates [28, 43], though the integrated healthcare system reported a screening rate of 60% [44]. The low screening rate is attributed to the difficulties in performing OGTT, patients’ fear of being diagnosed with diabetes, and lack of screening at postpartum visits [45]. Existing diagnostic tests, such as FPG, 2-hour plasma glucose, and HbA1c, have performance limitations in identifying high-risk populations [46]. Women with a history of GDM are seven times more likely to develop T2DM in later life [29]. OGTT is currently used to assess the risk of T2DM in women with a history of GDM, but it has many limitations. There is a lack of accurate risk stratification methods for postpartum T2DM. Risk prediction models may aid in clinical decision-making for patient management [47]. In recent years, ML has garnered much attention [48]. ML, a subset of artificial intelligence, focuses on developing computer systems that can identify potential patterns in training data to perform classification and prediction tasks on new data. ML integrates tools from statistics, data mining, and optimization to create models [49]. It has potentially ideal value and exhibits superior performance to OGTT.

Among ML algorithms (LR, DT, NB, Cox, and RF) in the included studies, Cox and DT were the most utilized. No significant difference was observed in the C-statistics of these ML algorithms. LR exhibited inferior predictive accuracy to other algorithms. Nevertheless, the lack of significant differences in their predictive performance suggests that further investigation is warranted to figure out more effective predictors in future analyses.

Variables

Variables included lipolytic metabolites, independent predictors of T2DM, and biomarkers of metabolic diseases. LR and DT were predominantly used among ML models. Two eligible studies had the highest C-statistics. ML optimization in a DT format revealed a T2DM predictive signature [27] with a C-statistic of 0.92 and an accuracy of 91%. The accuracy of this predictive signature was not affected by sensitivity (87%) or specificity (93%) and was more accurate than clinical diagnoses. Circulating microRNAs [50] are well-established biomarkers for metabolic diseases. The inclusion of circulating miR-369-3p measured at 12 weeks postpartum significantly increased the C-statistic from 0.83 (95% CI 0.68, 0.97) to 0.92 (95% CI: 0.84, 1.00) for predicting future T2DM in women with previous GDM [35]. Incorporating lipid metabolites and biomarkers of metabolic diseases into the prediction model for the progression of GDM to T2DM provides greater prediction accuracy for future risk [51].

Advantages and limitations

This paper first summarized the predictive accuracy of ML for the risk of T2DM in women with a history of GDM and provided evidence for the development and updating of subsequent clinical prediction tools. Nevertheless, there are some limitations. First, a few studies and models were included in the present paper after a comprehensive and systematic review. Second, many included studies were from single centers and lacked independent external validation. Only a few of the studies included had external validation, which may limit the interpretation of our results. Third, some studies included had fewer cases. Fourth, the follow-up duration of the studies included varied significantly. Due to the small number of models included, we did not discuss the differences in follow-up durations in depth under the same model. Heterogeneity may be introduced due to varied models, outcome measures, and follow-up durations, which is a limitation of our study.

Conclusion

ML has favorable predictive accuracy for the progression to T2DM in women with a history of GDM and can be used as a potential predictive tool. However, this conclusion is based on limited evidence. Therefore, real-time prediction models with broader applicability should be developed for GDM patients of different ethnic backgrounds. Individualized health interventions and lifestyle recommendations for high-risk populations will be beneficial to reduce the risk of T2DM.

Data availability

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

References

  1. Wang J, Zheng J, Shi W, Du N, Xu X, Zhang Y, et al. Dysbiosis of maternal and neonatal microbiota associated with gestational diabetes mellitus. Gut. 2018;67(9):1614–25. https://doi.org/10.1136/gutjnl-2018-315988.

    Article  PubMed  Google Scholar 

  2. Saravanan P. Gestational diabetes: opportunities for improving maternal and child health. Lancet Diabetes Endocrinol. 2020;8(9):793–800. https://doi.org/10.1016/s2213-8587(20)30161-3.

    Article  PubMed  Google Scholar 

  3. Ravaut M, Harish V, Sadeghi H, Leung KK, Volkovs M, Kornas K, et al. Development and validation of a Machine Learning Model Using Administrative Health Data to predict onset of type 2 diabetes. JAMA Netw Open. 2021;4(5):e2111315. https://doi.org/10.1001/jamanetworkopen.2021.11315.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Vounzoulaki E, Khunti K, Abner SC, Tan BK, Davies MJ, Gillies CL. Progression to type 2 diabetes in women with a known history of gestational diabetes: systematic review and meta-analysis. BMJ. 2020;369:m1361. https://doi.org/10.1136/bmj.m1361.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Ratner RE, Christophi CA, Metzger BE, Dabelea D, Bennett PH, Pi-Sunyer X, et al. Prevention of diabetes in women with a history of gestational diabetes: effects of metformin and lifestyle interventions. J Clin Endocrinol Metab. 2008;93(12):4774–9. https://doi.org/10.1210/jc.2008-0772.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Green JB. Cardiovascular consequences of Gestational Diabetes. Circulation. 2021;143(10):988–90. https://doi.org/10.1161/circulationaha.120.052995.

    Article  PubMed  Google Scholar 

  7. Zhu Y, Zhang C. Prevalence of gestational diabetes and risk of progression to type 2 diabetes: a global perspective. Curr Diab Rep. 2016;16(1):7. https://doi.org/10.1007/s11892-015-0699-x.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Kim C, Berger DK, Chamany S. Recurrence of gestational diabetes mellitus: a systematic review. Diabetes Care. 2007;30(5):1314–9. https://doi.org/10.2337/dc06-2517.

    Article  PubMed  Google Scholar 

  9. Khosla S, Samakkarnthai P, Monroe DG, Farr JN. Update on the pathogenesis and treatment of skeletal fragility in type 2 diabetes mellitus. Nat Rev Endocrinol. 2021;17(11):685–97. https://doi.org/10.1038/s41574-021-00555-5.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Segar MW, Vaduganathan M, Patel KV, McGuire DK, Butler J, Fonarow GC, et al. Machine learning to predict the risk of Incident Heart failure hospitalization among patients with diabetes: the WATCH-DM risk score. Diabetes Care. 2019;42(12):2298–306. https://doi.org/10.2337/dc19-0587.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Carson MP, Ananth CV, Gyamfi-Bannerman C, Smulian J, Wapner RJ. Postpartum Testing to Detect Persistent Dysglycemia in Women with Gestational Diabetes Mellitus. Obstet Gynecol. 2018;132(1):193–8. https://doi.org/10.1097/aog.0000000000002687.

    Article  PubMed  Google Scholar 

  12. Hod M, Kapur A, McIntyre HD. Evidence in support of the International Association of Diabetes in pregnancy study groups’ criteria for diagnosing gestational diabetes mellitus worldwide in 2019. Am J Obstet Gynecol. 2019;221(2):109–16. https://doi.org/10.1016/j.ajog.2019.01.206.

    Article  PubMed  Google Scholar 

  13. Chen X, Zhang Y, Chen H, Jiang Y, Wang Y, Wang D, et al. Association of Maternal Folate and Vitamin B(12) in early pregnancy with gestational diabetes Mellitus: a prospective cohort study. Diabetes Care. 2021;44(1):217–23. https://doi.org/10.2337/dc20-1607.

    Article  PubMed  Google Scholar 

  14. Salvia MG, Quatromoni PA. Behavioral approaches to nutrition and eating patterns for managing type 2 diabetes: a review. Am J Med Open. 2023;9:100034. https://doi.org/10.1016/j.ajmo.2023.100034.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Kirwan JP, Sacks J, Nieuwoudt S. The essential role of exercise in the management of type 2 diabetes. Cleve Clin J Med. 2017;84(7 Suppl 1):S15–21. https://doi.org/10.3949/ccjm.84.s1.03.

    Article  PubMed  PubMed Central  Google Scholar 

  16. AlKhudidi FH, Alsulaimani AI, Alharthi AH, Alrumaym AH, Alharthi EK, Altalhi WA, et al. Awareness of type 2 Diabetic patients about the importance of Exercise and Diet on Diabetes Type 2 in the Western Region of Saudi Arabia. Mater Sociomed. 2021;33(4):276–81. https://doi.org/10.5455/msm.2021.33.276-281.

    Article  PubMed  PubMed Central  Google Scholar 

  17. Zhang Z, Yang L, Han W, Wu Y, Zhang L, Gao C, et al. Machine learning prediction models for gestational diabetes Mellitus: Meta-analysis. J Med Internet Res. 2022;24(3):e26634. https://doi.org/10.2196/26634.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Andaur Navarro CL, Damen JAA, Takada T, Nijman SWJ, Dhiman P, Ma J, et al. Risk of bias in studies on prediction models developed using supervised machine learning techniques: systematic review. BMJ. 2021;375:n2281. https://doi.org/10.1136/bmj.n2281.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Bedrikovetski S, Dudi-Venkata NN, Kroon HM, Seow W, Vather R, Carneiro G, et al. Artificial intelligence for pre-operative lymph node staging in colorectal cancer: a systematic review and meta-analysis. BMC Cancer. 2021;21(1):1058. https://doi.org/10.1186/s12885-021-08773-w.

    Article  PubMed  PubMed Central  Google Scholar 

  20. Bedrikovetski S, Dudi-Venkata NN, Maicas G, Kroon HM, Seow W, Carneiro G, et al. Artificial intelligence for the diagnosis of lymph node metastases in patients with abdominopelvic malignancy: a systematic review and meta-analysis. Artif Intell Med. 2021;113:102022. https://doi.org/10.1016/j.artmed.2021.102022.

    Article  PubMed  Google Scholar 

  21. 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. https://doi.org/10.1186/s13073-021-00968-x.

    Article  PubMed  PubMed Central  Google Scholar 

  22. Yang Y, Xu L, Sun L, Zhang P, Farid SS. Machine learning application in personalised lung cancer recurrence and survivability prediction. Computational and Structural Biotechnology Journal. 2022; 20:1811–20. https://doi.org/10.1016/j.csbj.2022.03.035

  23. Liang X, Yu X, Gao T. Machine learning with magnetic resonance imaging for prediction of response to neoadjuvant chemotherapy in breast cancer: a systematic review and meta-analysis. Eur J Radiol. 2022;150:110247. https://doi.org/10.1016/j.ejrad.2022.110247.

    Article  PubMed  Google Scholar 

  24. Vieira S, Liang X, Guiomar R, Mechelli A. Can we predict who will benefit from cognitive-behavioural therapy? A systematic review and meta-analysis of machine learning studies. Clin Psychol Rev. 2022;97:102193. https://doi.org/10.1016/j.cpr.2022.102193.

    Article  PubMed  Google Scholar 

  25. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JP, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009;6(7):e1000100. https://doi.org/10.1371/journal.pmed.1000100.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Debray TP, Damen JA, Riley RD, Snell K, Reitsma JB, Hooft L, et al. A framework for meta-analysis of prediction model studies with binary and time-to-event outcomes. Stat Methods Med Res. 2019;28(9):2768–86. https://doi.org/10.1177/0962280218785504.

    Article  PubMed  Google Scholar 

  27. Khan SR, Mohan H, Liu Y, Batchuluun B, Gohil H, Al Rijjal D, et al. The discovery of novel predictive biomarkers and early-stage pathophysiology for the transition from gestational diabetes to type 2 diabetes. Diabetologia. 2019;62(4):687–703. https://doi.org/10.1007/s00125-018-4800-2.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Allalou A, Nalla A, Prentice KJ, Liu Y, Zhang M, Dai FF, et al. A predictive metabolic signature for the Transition from Gestational Diabetes Mellitus to type 2 diabetes. Diabetes. 2016;65(9):2529–39. https://doi.org/10.2337/db15-1720.

    Article  PubMed  PubMed Central  Google Scholar 

  29. Li W, Leng J, Liu H, Zhang S, Wang L, Hu G, et al. Nomograms for incident risk of post-partum type 2 diabetes in Chinese women with prior gestational diabetes mellitus. Clin Endocrinol. 2019;90(3):417–24. https://doi.org/10.1111/cen.13863.

    Article  Google Scholar 

  30. Lai M, Liu Y, Ronnett GV, Wu A, Cox BJ, Dai FF, et al. Amino acid and lipid metabolism in post-gestational diabetes and progression to type 2 diabetes: a metabolic profiling study. PLoS Med. 2020;17(5):e1003112. https://doi.org/10.1371/journal.pmed.1003112.

    Article  PubMed  PubMed Central  Google Scholar 

  31. Lin HC, Su CT, Wang PC. An application of artificial immune recognition system for prediction of diabetes following gestational diabetes. J Med Syst. 2011;35(3):283–9. https://doi.org/10.1007/s10916-009-9364-8.

    Article  PubMed  Google Scholar 

  32. Köhler M, Ziegler AG, Beyerlein A. Development of a simple tool to predict the risk of postpartum diabetes in women with gestational diabetes mellitus. Acta Diabetol. 2016;53(3):433–7. https://doi.org/10.1007/s00592-015-0814-0.

    Article  PubMed  Google Scholar 

  33. Periyathambi N, Parkhi D, Ghebremichael-Weldeselassie Y, Patel V, Sukumar N, Siddharthan R, et al. Machine learning prediction of non-attendance to postpartum glucose screening and subsequent risk of type 2 diabetes following gestational diabetes. PLoS ONE. 2022;17(3):e0264648. https://doi.org/10.1371/journal.pone.0264648.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Man B, Schwartz A, Pugach O, Xia Y, Gerber B. A clinical diabetes risk prediction model for prediabetic women with prior gestational diabetes. PLoS ONE. 2021;16(6):e0252501. https://doi.org/10.1371/journal.pone.0252501.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Joglekar MV, Wong WKM, Ema FK, Georgiou HM, Shub A, Hardikar AA, et al. Postpartum circulating microRNA enhances prediction of future type 2 diabetes in women with previous gestational diabetes. Diabetologia. 2021;64(7):1516–26. https://doi.org/10.1007/s00125-021-05429-z.

    Article  PubMed  Google Scholar 

  36. Ilari L, Piersanti A, Göbl C, Burattini L, Kautzky-Willer A, Tura A, et al. Unraveling the factors determining development of type 2 diabetes in Women with a history of gestational diabetes Mellitus through Machine-Learning techniques. Front Physiol. 2022;13:789219. https://doi.org/10.3389/fphys.2022.789219.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Houri O, Gil Y, Chen R, Wiznitzer A, Hochberg A, Hadar E et al. Prediction of type 2 diabetes Mellitus according to glucose metabolism patterns in pregnancy using a Novel Machine Learning Algorithm. J J Med Biol Eng. 2022.

  38. Lappas M, Mundra PA, Wong G, Huynh K, Jinks D, Georgiou HM, et al. The prediction of type 2 diabetes in women with previous gestational diabetes mellitus using lipidomics. Diabetologia. 2015;58(7):1436–42. https://doi.org/10.1007/s00125-015-3587-7.

    Article  PubMed  Google Scholar 

  39. Belsti Y, Moran LJ, Goldstein R, Mousa A, Cooray SD, Baker S, et al. Development of a risk prediction model for postpartum onset of type 2 diabetes mellitus, following gestational diabetes; the lifestyle InterVention in gestational diabetes (LIVING) study. Clin Nutr. 2024;43(8):1728–35. https://doi.org/10.1016/j.clnu.2024.06.006.

    Article  PubMed  Google Scholar 

  40. Jiang W, Wang J, Shen X, Lu W, Wang Y, Li W, et al. Establishment and validation of a risk Prediction Model for Early Diabetic kidney Disease based on a Systematic Review and Meta-analysis of 20 cohorts. Diabetes Care. 2020;43(4):925–33. https://doi.org/10.2337/dc19-1897.

    Article  PubMed  Google Scholar 

  41. 12. Management of Diabetes in Pregnancy. Diabetes Care. 2016;39(Suppl 1):S94–8. https://doi.org/10.2337/dc16-S015.

    Article  Google Scholar 

  42. Standards of medical care in diabetes–2014. Diabetes Care. 2014;37(Suppl 1):S14–80. https://doi.org/10.2337/dc14-S014.

    Article  Google Scholar 

  43. Blatt AJ, Nakamoto JM, Kaufman HW. Gaps in diabetes screening during pregnancy and postpartum. Obstet Gynecol. 2011;117(1):61–8. https://doi.org/10.1097/AOG.0b013e3181fe424b.

    Article  PubMed  Google Scholar 

  44. Ferrara A. Increasing prevalence of gestational diabetes mellitus: a public health perspective. Diabetes Care. 2007;30(Suppl 2):S141–6. https://doi.org/10.2337/dc07-s206.

    Article  PubMed  Google Scholar 

  45. Russell MA, Phipps MG, Olson CL, Welch HG, Carpenter MW. Rates of postpartum glucose testing after gestational diabetes mellitus. Obstet Gynecol. 2006;108(6):1456–62. https://doi.org/10.1097/01.Aog.0000245446.85868.73.

    Article  PubMed  Google Scholar 

  46. Jagannathan R, Neves JS, Dorcely B, Chung ST, Tamura K, Rhee M, et al. The oral glucose tolerance test: 100 years later. Diabetes Metab Syndr Obes. 2020;13:3787–805. https://doi.org/10.2147/dmso.S246062.

    Article  PubMed  PubMed Central  Google Scholar 

  47. Collins GS, Mallett S, Omar O, Yu LM. Developing risk prediction models for type 2 diabetes: a systematic review of methodology and reporting. BMC Med. 2011;9:103. https://doi.org/10.1186/1741-7015-9-103.

    Article  PubMed  PubMed Central  Google Scholar 

  48. Kodama S, Fujihara K, Horikawa C, Kitazawa M, Iwanaga M, Kato K, et al. Predictive ability of current machine learning algorithms for type 2 diabetes mellitus: a meta-analysis. J Diabetes Investig. 2022;13(5):900–8. https://doi.org/10.1111/jdi.13736.

    Article  PubMed  PubMed Central  Google Scholar 

  49. Fregoso-Aparicio L, Noguez J, Montesinos L, García-García JA. Machine learning and deep learning predictive models for type 2 diabetes: a systematic review. Diabetol Metab Syndr. 2021;13(1):148. https://doi.org/10.1186/s13098-021-00767-9.

    Article  PubMed  PubMed Central  Google Scholar 

  50. Capobianco E. Systems and precision medicine approaches to diabetes heterogeneity: a Big Data perspective. Clin Transl Med. 2017;6(1):23. https://doi.org/10.1186/s40169-017-0155-4.

    Article  PubMed  PubMed Central  Google Scholar 

  51. Scirica BM. Use of biomarkers in Predicting the Onset, Monitoring the Progression, and risk stratification for patients with type 2 diabetes Mellitus. Clin Chem. 2017;63(1):186–95. https://doi.org/10.1373/clinchem.2016.255539.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

Not applicable.

Funding

Quality of life care research fund of Chinese Association For Life Care (HL20210145).

Author information

Authors and Affiliations

  1. Department of Endocrinology and Metabolic Diseases, The Affiliated Hospital of Medical College Qingdao University, Qingdao, Shandong, 266003, China

    Meng Zhao, Zhixin Yao, Yan Zhang, Lidan Ma, Wenquan Pang & Yijun Xu

  2. Department of Emergency Pediatric, The Affiliated Hospital of Medical College Qingdao University, Qingdao, Shandong, 266003, China

    Shuyin Ma

  3. Department of Nursing, The Affiliated Hospital of Medical College Qingdao University, Qingdao, Shandong, 266003, China

    Lili Wei

Authors
  1. Meng Zhao
    View author publications

    Search author on:PubMed Google Scholar

  2. Zhixin Yao
    View author publications

    Search author on:PubMed Google Scholar

  3. Yan Zhang
    View author publications

    Search author on:PubMed Google Scholar

  4. Lidan Ma
    View author publications

    Search author on:PubMed Google Scholar

  5. Wenquan Pang
    View author publications

    Search author on:PubMed Google Scholar

  6. Shuyin Ma
    View author publications

    Search author on:PubMed Google Scholar

  7. Yijun Xu
    View author publications

    Search author on:PubMed Google Scholar

  8. Lili Wei
    View author publications

    Search author on:PubMed Google Scholar

Contributions

Yijun Xu, Lili Wei, and Meng Zhao conceived the idea of this study; Meng Zhao and Lili Wei screened the literature and extracted the data; Meng Zhao conducted statistical analysis; Lidan Ma supervised the analysis. Meng Zhao, Zhixin Yao, Yan Zhang, Wenquan Pang, and Shuyin Ma explained the survey results; Meng Zhao drafted the manuscript. Yijun Xu, Lili Wei, Meng Zhao, and Lidan Ma critically reviewed the manuscript, and Meng Zhao revised the manuscript for final submission. All authors have approved the final draft of the manuscript. Yijun Xu is the guarantor. Meng Zhao takes full responsibility for the study and has the right to access data and decide the publication. The corresponding author certified that all listed authors met the author criteria, and no other authors who met the criteria were omitted.

Corresponding authors

Correspondence to Yijun Xu or Lili Wei.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1: Table S1 Search strategy

Supplementary Material 2

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhao, M., Yao, Z., Zhang, Y. et al. Predictive value of machine learning for the progression of gestational diabetes mellitus to type 2 diabetes: a systematic review and meta-analysis. BMC Med Inform Decis Mak 25, 18 (2025). https://doi.org/10.1186/s12911-024-02848-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12911-024-02848-x

Keywords

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us