Today, an estimated 10 percent of the U.S. population is living with diabetes, and according to the CDC, one in three Americans could develop diabetes by 2050. These patients with diabetes are at higher risk for comorbidities like retinopathy — which can cause blindness — and nephropathy — which can cause kidney failure — as well as stroke and cardiovascular events. For the more than 30 million people diagnosed with diabetes in 2017, medical costs totaled $327 billion, per the American Diabetes Association.
Prediabetes, the precursor to diabetes, can be managed or even reversed through early intervention and action. Prediabetes can be identified early at primary care visits, through screening blood tests (such as blood glucose levels or HbA1c). In addition, risk surveys for prediabetes are readily available through the ADA website. These surveys take into account individual patient characteristics, such as age, sex, weight, family history and other existing medical problems.
Patients who are identified at the prediabetes stage can be referred to diabetes prevention programs, which offer support and education around needed lifestyle changes like healthy eating, exercise and weight management. By adopting these changes, individuals with prediabetes can delay or prevent the onset of diabetes.
Once prediabetes progresses to diabetes, however, it is much harder to control and can cause serious complications. Unfortunately, as the condition of prediabetes is often silent, 90 percent of those who have it today do not know it. This necessitates awareness and screening campaigns for both affected patients and their clinicians. For clinicians to quickly and effectively identify at-risk patients, they require both knowledge and intuitive tools to bring them the correct information within their workflows, enabling fully informed decision-making at the point of care. These decisions can involve both prescribing medication and referrals to lifestyle and educational programs.
Using data to identify and predict at-risk populations
To optimize treatment pathways for a precision medicine approach to care for persons with diabetes or prediabetes, we need to better understand how conditions are impacted by numerous factors. These include environment and prior medical history. At Allscripts, we contribute to this knowledge by leveraging our big data framework to derive insights on prediabetes and diabetes progression, generating predictive models based on today’s most advanced analytics tools.
In 2013, we created rules consistent with the CDC definitions for prediabetes, including patients with HbA1c levels from 5.7 percent to 6.4 percent, and factoring in body mass index, age and race. Applying these rules to our own EHR data, we found more than 3.5 million patients who fit the criteria and were at risk for diabetes.
As we followed prediabetes patient pathways over a four-year period, we found in some geographic areas, almost 80 percent of them went on to develop diabetes. This unexpected finding was much higher than the CDC’s estimation of one in three patients. Our ability to conduct this large-scale analysis and find such significant progression to diabetes suggests this problem may be grossly underestimated, especially with more than 80 million people at risk for diabetes in the U.S. today. Furthermore, while these results demonstrate the power of predicting at-risk populations, they reiterate the need for resources that empower clinicians, and patients to respond to and mitigate these risks at the individual level.
Clinician burden and approach to personalized treatment at the point of care
Individual physicians often have limited time with patients, and diabetes is frequently only one of many chronic diseases they manage. One way to reduce clinician burden is to deliver standards of care directly within the workflows they use every day. Each year, the ADA updates its standards-of-care document for clinicians, so they can best care for patients with prediabetes or diabetes. The document includes recommendations for referrals to specialist care for complex diabetes, self-management programs or medications to optimize management of diabetes.
The Allscripts EHR diabetes package was built to deliver clinical decision support to the point of care, and translates the ADA’s most recent standards of care documentation into EHR-friendly clinical workflows, using a powerful rules engine to deliver relevant algorithms to clinicians. Through CDS, we can ensure patients across a population receive a consistent standard of care, overcoming the variations in clinical practices that can lead to negative outcomes.
The future
Delivering standards of care is just the first step to enabling the best possible care. As additional advancements are made in research, we hope to enrich clinicians by giving them disease progression patterns in demographically similar patients, taking into account social and environmental risk factors, which may not appear in the health data. This information can enable optimized treatment plans. Patients, in turn, would receive personalized resources needed to take charge of their health and make the most effective changes given their own unique, individual circumstances.
As a clinician, I understand the value of up-to-date and data-driven CDS, especially those with complex chronic diseases like diabetes. Providing tools, based on broadly sourced population data and intelligent predictive models, enable clinicians to maintain the highest standard of care delivery while reducing burnout, and ensuring patients receive quality care. Incorporating predictive models into CDS, can enhance care plans leading to the most optimal health outcomes.