Diabetes Epidemic Continues to Grow, Stressing the Need for Earlier CKD Diagnosis

Steven Coca, DO, outlines the importance of early identification of kidney disease in patients with diabetes and the potential to use artificial intelligence for this purpose.

I began practicing and treating patients with chronic kidney disease (CKD) more than 12 years ago. As my career in nephrology has progressed, CKD has also progressed and is one of the most critical public health challenges in the United States and worldwide. The statistics are daunting: currently, there are 850 million individuals with CKD worldwide,1 with 37 million in the United States.2

The growing prevalence of CKD can be, in part, attributed to the steady increase in the number of individuals developing diabetes, one of the primary causes of CKD.3 It has been reported that approximately 12 million individuals in the United States have diabetic kidney disease (DKD)4,5 and 38% of end-stage kidney disease in the United States is caused by diabetes;2 more than any other cause, including high blood pressure. Currently, 44% of new cases of CKD are caused by diabetes,6 and this number has continued to grow. According to study results published in Seminars in Nephrology, as diabetes becomes more prevalent, it will cause the number of cases of renal failure to quadruple in the coming decades.7

The Growing Toll of CKD on Patients and the US Healthcare System

As the CKD epidemic continues to grow, it is having a profound impact on patients, families, hospitals, and the US healthcare system overall. But challenges remain in both diagnosis and treatment. According to a study from the National Kidney Foundation, 46% of Americans — including a number of individuals who had been diagnosed with diabetes — were not aware that having diabetes puts an individual at greater risk for kidney failure, and 31% were not aware that it puts an individual at increased risk of developing kidney disease.8

In reality, DKD is the main cause of kidney failure in 45% of patients who develop end-stage renal disease (ESRD).9 Many of these individuals are not aware of the actual status of their reduced kidney function. Thus, many patients progress to kidney failure in an unplanned manner — up to 63% of patients with existing kidney disease initiate dialysis in a “crash” or unexpected fashion.10 In addition, many primary care providers are not aware of or do not act on rapidly declining kidney function until it is too late, emphasizing the need for improved, reliable risk assessment tools in the predialysis setting.

Related Articles

Along with the effect on patients, CKD and DKD have a tremendous effect on the nation’s healthcare system. Approximately 20% of dollars in traditional Medicare — $114 billion in 2018 — are spent on Americans with kidney disease.11,12 The United States Renal Data System (USRDS) has stated that kidney disease is a cost multiplier, as its accompanying conditions, such as diabetes and heart failure, increase the cost of caring for individuals with CKD. The organization asserts that there could be substantial cost savings if these diseases are detected and treated early, which would prevent further complications.10

Given the increasing prevalence of diabetes and CKD, as well as the effect CKD has and will continue to have on patients and the healthcare system, it is more important now than ever for healthcare providers to have the ability to detect kidney disease at its earliest stages. Early and accurate risk stratification of individuals likely to experience rapid kidney function decline or kidney failure is critical for improving patient outcomes, clinical resource allocation, and cost control.

How AI Can Make an Impact on CKD and DKD

On July 10, 2019, President Trump launched the Advancing American Kidney Health Initiative directing the Department of Health and Human Services (HHS) to take bold action to transform how kidney disease is prevented, diagnosed, and treated in the next decade.13 Administration officials from HHS, the Centers for Medicare and Medicaid Services, the US Food and Drug Administration, the US Centers for Disease Control and Prevention, and other federal agencies have defined 3 goals for improving kidney health. The first goal is to reduce the number of Americans developing ESRD by 25% by 2030.13

One of the most promising methods for enabling healthcare providers to identify and properly stratify risk for patients likely to experience fast-progressing kidney disease or advance to kidney failure is artificial intelligence (AI) combined with electronic health records and biomarkers.

Using machine learning algorithms to analyze electronic health record information and proven predictive blood-based biomarkers associated with kidney disease, researchers have created a predictive model to help identify patients with CKD with high risk of progressing to dialysis and transplant and patients who are at a lower risk for progression.14 Used in conjunction with clinical evaluation to match appropriate management and treatment strategies that can delay or prevent progression to kidney failure, this application of AI for CKD holds great promise as it will enable healthcare providers to better stratify patients and determine which patients must be seen by a specialist.

Identifying at-risk patients earlier in the disease cycle will allow actions to be taken to slow or arrest kidney function decline, including but not limited to nephrology referral, dietary education, better blood pressure control, and the use of new kidney-protective agents such as sodium-glucose cotransporter 2 inhibitors. Patients with DKD found to be at low risk for rapid kidney function decline will avoid potentially unnecessary specialist referral and medical interventions. More accurate decisions about referral, monitoring, and treatment will have a positive impact on long-term patient outcomes, reducing the percentage of patients with DKD who progress to ESRD, dialysis, and kidney transplant.

In addition, identifying at-risk patients earlier can support consultative care between nephrologists and patients when dialysis cannot be avoided, increasing the utilization of home dialysis vs center-driven hemodialysis in the United States.

Looking Forward

There is no doubt that AI holds great potential for healthcare, and specifically for CKD.

Coupled with dedicated clinical care, AI can help us focus care efforts on the right patients at the right time, to address potential negative events before they happen and help improve patient outcomes.

We are just now beginning to see the first wave of AI-enabled diagnostic technologies for the nephrology and endocrinology market, and so far the results are very promising. I expect that this interest will continue to grow in the years to come as healthcare providers begin to see the results of AI’s application.

Disclosure: Steven Coca, DO, is associate chair of research for the department of medicine at Mount Sinai, and cofounder of RenalytixAI, a developer of AI-enabled diagnostics for kidney disease.

Follow @EndoAdvisor


1. The hidden epidemic: Worldwide, over 850 million people suffer from kidney diseases [news release]. Weimar, Germany: The International Society of Nephrology; June 27, 2018. https://www.theisn.org/images/ASN_PR_20180627_Final6.26.18Press_E.pdf. Accessed November 22, 2019.

2. Centers for Disease Control and Prevention. Chronic kidney disease in the United States, 2019. https://www.cdc.gov/kidneydisease/publications-resources/2019-national-facts.html. Published March 11, 2019. Accessed November 26, 2019.

3. National Kidney Foundation. Kidney disease: the basics. https://www.kidney.org/news/newsroom/factsheets/KidneyDiseaseBasics. Published September 4, 2019. Accessed November 22, 2019.

4. Alicic RZ, Rooney MT, Tuttle KR. Diabetic kidney disease: challenges, progress, and possibilities. Clin J Am Soc Nephrol. 2017;12(12):2032-2045.

5. Centers for Disease Control and Prevention. Type 2 diabetes. https://www.cdc.gov/diabetes/basics/type2.html. Published May 30, 2019. Accessed January 20, 2020.

6. National Kidney Foundation. Diabetes and chronic kidney disease. https://www.kidney.org/news/newsroom/factsheets/Diabetes-And-CKD. Published January 27, 2016. Accessed December 20, 2019.

7. Breyer MD, Susztak K. Developing treatments for chronic kidney disease in the 21st century. Semin Nephrol. 2016;36(6):436-447.

8. New Harris Poll Shows Many Americans Don’t Know About Kidney Disease Risk [news release]. New York, NY: National Kidney Foundation; November 14, 2019. https://www.kidney.org/news/new-harris-poll-shows-many-americans-don’t-know-about-kidney-disease-risk. Accessed November 26, 2019.

9. Ghaderian SB, Hayati F, Shayanpour S, Beladi Mousavi SS. Diabetes and end-stage renal disease; a review article on new concepts. J Renal Inj Prev. 2015;4(2):28-33.

10. Molnar AO, Hiremath S, Brown PA, Akbari A. Risk factors for unplanned and crash dialysis starts: a protocol for a systematic review and meta-analysis. Syst Rev. 2016;5:117.

11. United States Renal Data System. Annual data report highlights. https://www.usrds.org/adrhighlights.aspx. Published 2018. Accessed December 10, 2019.

12. United States Renal Data System. USRDS Annual Data Report | Volume 1 – CKD in the United States. Chapter 6: healthcare expenditures for persons with CKD. https://www.usrds.org/2017/view/v1_06.aspx. Published 2017. Accessed January 20, 2020.

13. Azar II AM; United States Department of Health and Human Services. Advancing American kidney health. https://aspe.hhs.gov/system/files/pdf/262046/AdvancingAmericanKidneyHealth.pdf. Published July 10, 2019. Accessed January 16, 2020.

14. Nadkarni GN, Fleming F, McCullough JR, et al. Prediction of rapid kidney function decline using machine learning combining blood biomarkers and electronic health record data [published online March 28, 2019]. bioRxiv. doi:10.1101/587774