For patients with chronic kidney disease (CKD), the diagnostic journey can be long and arduous. Unfortunately, many individuals are not diagnosed until later stages of the disease. As a result, patients often end up with irreversible kidney damage. However, this might soon change, reports Medical XPress. Researchers from Columbia University’s Vagelos College of Physicians and Surgeons recently created a medical algorithm designed to interpret electronic medical record data and identify early-stage CKD. See the full study information published in npj Digital Medicine.

Medical Algorithm

Earlier CKD detection can lead to improved treatment opportunities, slowed disease progression, and better patient outcomes. However, only 10% of patients with early-stage CKD know that they have it. Oftentimes, patients in the early stages do not have symptoms. In other cases, symptoms are mild enough that doctors do not prioritize them. Finally, the diagnostic process is complicated by diagnostic tests which are difficult to interpret when considering patient demographics and characteristics. Thus, researchers knew they needed to find a potential solution.

Altogether, this new technology uses an algorithm to assess blood and urine test results, and determine whether the data suggests early-stage CKD. To do this, the algorithm uses a mix of equations and machine learning, a system of artificial intelligence (AI) designed to make inferences with little human intervention. According to data, the algorithm identifies cases of CKD nearly as well as experienced nephrologists. In fact, the algorithm correctly ruled out CKD in 97% of controls and identified it in 95% of patients, when exploring 451 medical records.

While the algorithm can be complicated by a lack of blood or urine tests, researchers believe it can still be helpful in the diagnostic process. It can also be applied to a variety of other datasets, assisting researchers in other studies. For example, it can help determine whether CKD is associated with any other conditions, such as depression, or whether there are specific genes associated with the condition.

Chronic Kidney Disease (CKD)

Normally, chronic kidney disease (CKD), or chronic kidney failure, occurs following some disease or condition which negatively impacts the kidneys. For example, diabetes, high blood pressure, polycystic kidney disease, urinary tract obstruction, and interstitial nephritis may all play a role in the development of CKD. Ultimately, CKD is progressively worsening kidney damage. In later stages, fluids and other wastes accumulate in the body, as the kidneys are unable to filter these out of the blood. An estimated 26 million Americans have CKD. Symptoms include:

• Nausea and vomiting
• Appetite loss
• General fatigue and weakness
• Jaundice (yellowing of the skin and eyes)
• Unintended weight loss
• Liver or spleen enlargement
• Muscle cramping
• Liver diseases
• High blood pressure
• Shortness of breath and/or difficulty breathing
• Pruritus (intense and persistent itching)
• Swelling of the lower extremities
• Bloody urine
• Increased need for urination
• Anemia (low red blood cell count)
• Vitamin deficiencies