Abstract
A large proportion of cases with chronic conditions including diabetes or pre-diabetes, hypertension and dyslipidemia remain undiagnosed. To include reproductive factors (RF) might be able to improve current screening guidelines by providing extra effectiveness. The objective is to study the relationships between RFs and chronic conditions’ biomarkers. A cross-sectional study was conducted. Demographics, RFs and metabolic biomarkers were collected. The relationship of the metabolic biomarkers were shown by correlation analysis. Principal component analysis (PCA) and autoencoder were compared by cross-validation. The better one was adopted to extract a single marker, the general chronic condition (GCC), to represent the body’s chronic conditions. Multivariate linear regression was performed to explore the relationship between GCC and RFs. In total, 1,656 postmenopausal females were included. A multi-layer autoencoder outperformed PCA in the dimensionality reduction performance. The extracted variable by autoencoder, GCC, was verified to be representative of three chronic conditions (AUC for patoglycemia, hypertension and dyslipidemia were 0.844, 0.824 and 0.805 respectively). Linear regression showed that earlier age at menarche (OR = 0.9976) and shorter reproductive life span (OR = 0.9895) were associated with higher GCC. Autoencoder performed well in the dimensionality reduction of clinical metabolic biomarkers. Due to high accessibility and effectiveness, RFs have potential to be included in screening tools for general chronic conditions and could enhance current screening guidelines.
Similar content being viewed by others
Introduction
Type 2 diabetes mellitus (T2DM), hypertension and hyperlipidemia are chronic conditions that can result in severe complications1,2, including cardiovascular disease (CVD), the leading cause of death worldwide3,4. Unfortunately, a large number of patients with these conditions remain undiagnosed. Most updated literatures showed that in 2019 there are 50.1% (231.9 million) of diabetes patients still undiagnosed worldwide5. A large proportion of cases with hypertension6 and hyperlipidemia7 are also unware of their condition, particularly in low and middle income countries8.
Screening of those at risk of chronic conditions is of significance for both individuals and wider society, yet there are gaps in current practices. Early identification is generally based on commonly collected risk factors, such as age, gender, smoking status, body mass index (BMI) and family history. A number of societies and task forces have recommended various screening guidelines that consist of these risk factors9,10,11; however, there are growing concerns that such guidelines might be inadequate and inaccurate12,13,14,15. For example, the American Diabetes Association (ADA) and the US Preventive Services Task Force (USPSTF) guidelines have shown only a fair performance when externally validated12,13. Furthermore, a trial exploring the effectiveness of a population-based screening programme in the United Kingdom found that screening was not associated with a reduction in all-cause mortality over a median period of 9.6 years15. A number of commonly used screening functions have also been shown to be ineffective in population screening16.
To include novel or extra factors might help to identify high risk groups more accurately and has the potential to improve current screening guidelines for chronic conditions, in terms of both effectiveness and efficiency. Indeed, a growing body of studies has identified a strong relationship between women’s reproductive factors (RF) and chronic conditions. For example, early menarche has been found to be associated with an increased risk of T2DM17,18, obesity and insulin resistance19. Moreover, a retrospective study conducted in Europe showed that, after adjustment for confounding, early menopause and shorter reproductive life span was associated with T2DM20. A Japanese study also found a similar relationship regarding hypercholesterolemia21. Furthermore, the China Kadoorie Biobank study reported that Chinese women with late menopause (≥53 years) were 1.21 (95% CI: 1.03–1.42) times more likely to have T2DM than women with menopause at 46–52 years old (p < 0.0001)22. Another Chinese study also found that a higher number of live births was associated with hypertension and DM, and mediated by lifestyle and dyslipidemia55. High accessibility and low cost are outstanding advantages of RFs. Furthermore, there are studies that have shown that the validity and reproducibility of self-reported RFs are good56. Therefore, RFs have potential as screening tools for chronic conditions and could improve current screening guidelines.
A number of important limitations need to be considered. First, this is a cross-sectional study and hence it cannot infer causality. Second, this study only tested the possibility that RFs can be incorporated into a screening tool and did not give the actual sensitivity and specificity of RFs to screen for chronic conditions. In terms of further research, a structured screening tool should be developed and externally validated. Third, although not included in the current study, uric acid and HbA1c are also crucial biomarkers for chronic conditions and it is important that future research takes them into account. Last, interpretability is always a key concern when applying machine learning to medical data analysis. Many advanced methods have been proposed to unfold the black box of neuron networks. In future study, we hope to focus on this specific question and explore the GCC more comprehensively.
To conclude, autoencoder performed well in the dimensionality reduction of clinical biomarkers, demonstrating its potential in further medical data process. Women with earlier age at menarche and shorter reproductive life span are more likely to suffer from chronic conditions. Due to high accessibility and effectiveness, RFs show potential to be included in preliminary screening tools for general chronic conditions in clinical practice and could enhance current screening guidelines.
Data availability
The original data is not currently available online but can be requested in machine-readable format from the corresponding author on reasonable request.
References
Rahelic, D. 7th Edition of Idf Diabetes Atlas–Call for Immediate Action. Lijec Vjesn 138, 57–58 (2016).
Innovative care for chronic health conditions. Rev Panam Salud Publica 12, 71–74 (2002).
Dixon, J. & Dewar, S. The NHS plan. BMJ 321, 315–316 (2000).
Koh, H. K., Blakey, C. R. & Roper, A. Y. Healthy People 2020: a report card on the health of the nation. JAMA 311, 2475–2476, https://doi.org/10.1001/jama.2014.6446 (2014).
Saeedi, P. et al. Global and regional diabetes prevalence estimates for 2019 and projections for 2030 and 2045: Results from the International Diabetes Federation Diabetes Atlas, 9 edition. Diabetes research and clinical practice 157, 107843, https://doi.org/10.1016/j.diabres.2019.107843 (2019).
Mozaffarian. Heart Disease and Stroke Statistics-2016 Update: A Report From the American Heart Association (vol 133, pg e38, 2016). Circulation 133, E599–E599, https://doi.org/10.1161/CIR.0000000000000409 (2016).
Koyama, A. K., Bali, V., Yermilov, I. & Legorreta, A. P. Identification of Undiagnosed Hyperlipidemia: Do Work Site Screening Programs Work? American Journal of Health Promotion 32, 971–978, https://doi.org/10.1177/0890117116671537 (2018).
Who, A. J. W. H. O. Global brief on hypertension. (2013).
Abid, A., Ahmad, S. & Waheed, A. Screening for Type II Diabetes Mellitus in the United States: The Present and the Future. Clinical Medicine Insights: Endocrinology and Diabetes 9, https://doi.org/10.4137/CMED.S38247 (2016).
Gillman Matthew, W., Caspard, H., Lane, K., Forman John, P. & Rifas-Shiman Sheryl, L. Diabetes and lipid screening among patients in primary care: A cohort study. BMC Health Services Research 8, 25, https://doi.org/10.1186/1472-6963-8-25 (2008).
Screening for Lipid Disorders in Adults: Recommendation Statement. Am. Fam. Physician 80, 1273–1274 (2009).
Kai Mckeever, B. et al. Receipt of Glucose Testing and Performance of Two US Diabetes Screening Guidelines, 2007–2012. PLoS ONE 10, e0125249, https://doi.org/10.1371/journal.pone.0125249.
Sheehy, A. M. et al. Analysis of guidelines for screening diabetes mellitus in an ambulatory population. Mayo Clinic proceedings 85, 27–35, https://doi.org/10.4065/mcp.2009.0289 (2010).
Ochoa, P. S. et al. Identification of Previously Undiagnosed Diabetes and Prediabetes in the Inpatient Setting Using Risk Factor and Hemoglobin A1C Screening. Annals of Pharmacotherapy 48, 1434–1439, https://doi.org/10.1177/1060028014547383 (2014).
Nwaneri, C., Bowen-Jones, D. & Cooper, H. Screening for type 2 diabetes and population mortality over 10 years. Lancet (London, England) 381, 901, https://doi.org/10.1016/S0140-6736(13)60665-0 (2013).
Toscano, C. M. et al. Cost-effectiveness of a national population-based screening program for type 2 diabetes: the Brazil experience. Diabetol Metab Syndr 7, 95, https://doi.org/10.1186/s13098-015-0090-8 (2015).
He, C. et al. Age at menarche and risk of type 2 diabetes: results from 2 large prospective cohort studies. Am J Epidemiol 171, 334–344, https://doi.org/10.1093/aje/kwp372 (2010).
Muka, T. et al. Age at natural menopause and risk of type 2 diabetes: a prospective cohort study. Diabetologia 60, 1951–1960, https://doi.org/10.1007/s00125-017-4346-8 (2017).
Adair, L. S. & Gordon-Larsen, P. Maturational timing and overweight prevalence in US adolescent girls. American journal of public health 91, 642, https://doi.org/10.2105/AJPH.91.4.642 (2001).
Brand, J. S. et al. Age at menopause, reproductive life span, and type 2 diabetes risk: results from the EPIC-InterAct study. Diabetes care 36, 1012, https://doi.org/10.2337/dc12-1020 (2013).
Lee, J. S. et al. Independent association between age at natural menopause and hypercholesterolemia, hypertension, and diabetes mellitus: Japan nurses’ health study. Journal of Japan Atherosclerosis Society 20, 161–169 (2013).
Wang, M. et al. Age at natural menopause and risk of diabetes in adult women: Findings from the China Kadoorie Biobank study in the Zhejiang area. Journal of Diabetes Investigation 9, 762–768, https://doi.org/10.1111/jdi.12775 (2018).
Xu, B., Chen, Y., **ong, J. P., Lu, N. & Tan, X. Association of Female Reproductive Factors with Hypertension, Diabetes and LQTc in Chinese Women. Sci Rep, 7, https://doi.org/10.1038/srep42803 (2017).
Daniel, W. W. Biostatistics: a foundation for analysis in the health sciences. 9th ed. edn (John Wiley & Sons, 2009).
Feng, Y. et al. Prevalence of metabolic syndrome and its relation to body composition in a Chinese rural population. Obesity (Silver Spring) 14, 2089–2098, https://doi.org/10.1038/oby.2006.244 (2006).
Gu, D. et al. Prevalence of the metabolic syndrome and overweight among adults in China. Lancet 365, 1398–1405, https://doi.org/10.1016/S0140-6736(05)66375-1 (2005).
John A. RiceLes informations contenues dans cette page sont à usage strict de et ne doivent être utilisées ou copiées par un tiers.Powered by, and. Mathematical Statistics and Data Analysis, (Wadsworth, 1988).
Rumelhart, D. E., Hinton, G. E. & Williams, R. J. Learning representations by back-propagating errors. Nature 323, 533–536, https://doi.org/10.1038/323533a0 (1986).
Rubin, D. B. Multiple Imputation for Nonresponse in Surveys, https://doi.org/10.1002/9780470316696 (2008).
Kahn, R., Buse, J., Ferrannini, E. & Stern, M. The metabolic syndrome: time for a critical appraisal: joint statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes care 28, 2289–2304, https://doi.org/10.2337/diacare.28.9.2289 (2005).
DeBoer, M. D. & Gurka, M. J. Low sensitivity of the metabolic syndrome to identify adolescents with impaired glucose tolerance: An analysis of NHANES 1999-2010. Acta Veterinaria Scandinavica 13, 83, https://doi.org/10.1186/1475-2840-13-83 (2014).
Ford, E. S., Li, C. & Zhao, G. Prevalence and correlates of metabolic syndrome based on a harmonious definition among adults in the US. J Diabetes 2, 180–193, https://doi.org/10.1111/j.1753-0407.2010.00078.x (2010).
Beltran-Sanchez, H., Harhay, M. O., Harhay, M. M. & McElligott, S. Prevalence and trends of metabolic syndrome in the adult U.S. population, 1999–2010. J Am Coll Cardiol 62, 697–703, https://doi.org/10.1016/j.jacc.2013.05.064 (2013).
Obermeyer, Z. & Emanuel, E. J. Predicting the Future - Big Data, Machine Learning, and Clinical Medicine. N Engl J Med 375, 1216–1219 (2016).
Beam, A. L. & Kohane, I. S. Big Data and Machine Learning in Health Care. JAMA 319, 1317–1318, https://doi.org/10.1001/jama.2017.18391 (2018).
Hinton, G. E. & Salakhutdinov, R. R. Reducing the dimensionality of data with neural networks. Science 313, 504–507, https://doi.org/10.1126/science.1127647 (2006).
Majumdar, A., Gogna, A. & Ward, R. Semi-supervised Stacked Label Consistent Autoencoder for Reconstruction and Analysis of Biomedical Signals. IEEE Transactions on Biomedical Engineering PP, 1–1 (2016).
Plaut, E. From Principal Subspaces to Principal Components with Linear Autoencoders. (2018).
Wang, Y., Yao, H. & Zhao, S. Auto-encoder based dimensionality reduction. Neurocomputing 184, 232–242, https://doi.org/10.1016/j.neucom.2015.08.104 (2016).
Tan, C. & Eswaran, C. Reconstruction and recognition of face and digit images using autoencoders. Neural Comput & Applic 19, 1069–1079, https://doi.org/10.1007/s00521-010-0378-4 (2010).
Liu, M. et al. Association between reproductive variables and metabolic syndrome in chinese community elderly women. Archives of Gerontology & Geriatrics 63, 78–84 (2015).
Alonso, D. L. M. et al. Risk of ischemic stroke and lifetime estrogen exposure. Neurology 68, 33 (2007).
Lakshman, R. et al. Association between age at menarche and risk of diabetes in adults: results from the EPIC-Norfolk cohort study. Diabetologia 51, 781–786 (2008).
Mishra, G. D., Carrigan, G., Brown, W. J., Barnett, A. G. & Dobson, A. J. Short-term weight change and the incidence of diabetes in midlife: results from the Australian Longitudinal Study on Women’s Health. Diabetes care 30, 1418, https://doi.org/10.2337/dc06-2187 (2007).
Soriguer, S. F. et al. Type 2 diabetes mellitus and other cardiovascular risk factors are no more common during menopause: longitudinal study. Menopause 16, 817–821, https://doi.org/10.1097/GME.0b013e31819d4113 (2009).
Malacara, J. M., Huerta, R., Rivera, B., Esparza, S. & Fajardo, M. E. Menopause in normal and uncomplicated NIDDM women: physical and emotional symptoms and hormone profile. Maturitas 28, 35–45, https://doi.org/10.1016/S0378-5122(97)00051-0 (1997).
Cho, G. J. et al. Postmenopausal status according to years since menopause as an independent risk factor for the metabolic syndrome. Menopause-the Journal of the North American Menopause Society 15, 524–529 (2008).
Chakravarti, S. & Studd, J. Hormonal profiles after the menopause. British Medical Journal 2, 784–787 (1976).
Arca, M., Vega, G. L. & Grundy, S. M. Hypercholesterolemia in postmenopausal women: Metabolic defects and response to low-dose lovastatin. Jama 47, 87–88 (1994).
Gunderson, E. et al. Long-term plasma lipid changes associated with a first birth - The coronary artery risk development in young adults study. Am. J. Epidemiol. 159, 1028–1039, https://doi.org/10.1093/aje/kwh146 (2004).
Smith, D. E. et al. Longitudinal Changes in Adiposity Associated With Pregnancy: The CARDIA Study. JAMA 271, 1747–1751, https://doi.org/10.1001/jama.1994.03510460039030 (1994).
Laganà, A. S., Garzon, S., Casarin, J., Franchi, M. & Ghezzi, F. Inositol in Polycystic Ovary Syndrome: Restoring Fertility through a Pathophysiology-Based Approach. Trends in Endocrinology and Metabolism 29, 768–780, https://doi.org/10.1016/j.tem.2018.09.001 (2018).
Laganà, A. S., Vitale, S. G., Noventa, M. & Vitagliano, A. Current Management of Polycystic Ovary Syndrome: From Bench to Bedside. International Journal of Endocrinology 2018, https://doi.org/10.1155/2018/7234543 (2018).
Laganà, A. S. et al. Evidence-Based and Patient-Oriented Inositol Treatment in Polycystic Ovary Syndrome: Changing the Perspective of the Disease. International journal of endocrinology and metabolism 15, e43695, https://doi.org/10.5812/ijem.43695 (2017).
Bayati, M., Bhaskar, S. & Montanari, A. Statistical analysis of a low cost method for multiple disease prediction. Stat Methods Med Res 27, 2312–2328, https://doi.org/10.1177/0962280216680242 (2018).
Den Tonkelaar, I. Validity and reproducibility of self-reported age at menopause in women participating in the DOM-project. Maturitas 27, 117–123, https://doi.org/10.1016/S0378-5122(97)01122-5 (1997).
Acknowledgements
We are grateful for the efforts of all Research Assistants in the Chinese Academy of Science and the University of Hong Kong. Many thanks for your unremitting help and support. We would also like to express our appreciation to the patients in the primary care setting who consented to participate in this research. This study is supported by National Key R&D Program of the Ministry of Science and Technology of China (Grant No. 2016YFC1301602), National Natural Science Foundation of China (Grant Nos. 81871447 and 81661168015) and Shenzhen Science and Technology Innovation Commission (Grant No. JCYJ20160608153506088). The funding agencies played no role in the overall process of the research.
Author information
Authors and Affiliations
Contributions
J.L. and S.T. were involved in the whole process of this study, including study design, data collection, data process and results interpretation. S.T. and W.D. took charge of data analysis, model construction and testing, as well as the drafting of paper. K.C. and X.L. were responsible for the supervision of laboratory tests and also helped to revise the paper. L.B. contributed to the interpretation of the results and the revision of the paper. Prof. J.L. was the academic lead on this study and finalized the manuscript. All the authors have approved the final version of the publication, agree with the submission and accept responsibility for the paper’s validity.
Corresponding author
Ethics declarations
Competing interests
All the authors declare no competing interests. All have agreed to the submission to this journal and can confirm that the manuscript is not currently under submission in any other journal.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, 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 license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license 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 license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Tian, S., Dong, W., Chan, K.L. et al. Screening for chronic conditions with reproductive factors using a machine learning based approach. Sci Rep 10, 2848 (2020). https://doi.org/10.1038/s41598-020-59825-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-020-59825-3
- Springer Nature Limited