Many patients with early-stage chronic kidney disease (CKD) do not experience symptoms and go undiagnosed. This often leads to delays in care and inadequate prevention that can make their condition worse. In addition, undiagnosed CKD is having a significant financial impact on the healthcare system, as patients with end-stage kidney disease (ESKD) are more expensive to treat. 

Artificial intelligence (AI) models are helping to detect CKD when patients do not have visible symptoms. Learn about a study in Iraq that compared the accuracy of four different AI models in detecting CKD.*

What The Researchers Did

Researchers from the Baghdad Renal Clinic in Iraq collected 373,770 anonymized samples for their data set. The set was divided randomly into training data (70%) and test sets (30%).

Investigators used four AI matching algorithms to determine the most effective CKD detector, which included:

• Logistic regression
• Decision-forest
• Artificial neural network
• Jungle of decisions

What They Found

Decision forests were the highest-performing model, with 92.2% accuracy and a 92.1% completeness value, followed by neural networks (80.6% accuracy; 80.0% completeness) and decision jungle (75.4%% accuracy; 75.0%). The lowest performer was the logistic regression model, with 68.9% accuracy and a 68.9% completeness value.

What It Means

Due to the study’s limitations, such as the lack of significant data samples available because of restrictions in Iraq, researchers are calling for further analyses. 

“Thanks to the models, we can see how changing the characteristics affects the search for the target value with a simple change of column selection or improvements in the data. Furthermore, this methodology could apply to clinical data of other diseases’ and pathologies’ inaccurate medical diagnoses,” the investigators noted.

Ultimately, the AI models showed positive results and may be an accurate detector of CKD in the future. 

*Jeremias, S. (2022, January 19). Study: Which AI Model Is the Best at Detecting Early-Stage CKD? AJMC. https://www.ajmc.com/view/study-which-ai-model-is-the-best-at-detecting-early-stage-ckd-