<p>Digital health enables the development of optimal solutions for obtaining a correct diagnosis and treatment of diseases based on the medical data of the patient. One of the most important issues in this process is to be able to group patients dynamically by their symptoms in order to provide better and more relevant treatments. The new clustering architecture introduced in this paper, DCMHTS (Deep Clustering Model for Healthcare Triggered Systems), is a novel combination of the classification based on triggers and enhancements of feature extraction via some deep technics used for anomaly and health conditions detection. Differentiating from conventional clustering methods (e.g., K-Means and DBSCAN), which utilize static distance-based grouping to classify patients, DCMHTS uses adaptive triggers to dynamically classify patients according to symptom patterns. This model contains two main parts: (1) Detection of Anomaly Disease (DAD) which uses triggers to flag patients with symptom-based positive anomalies, and (2) Clustering Anomaly Disease (CAD) where the patients are assigned to the optimized clusters. Compared to other machine learning methods, our method has outperformed in terms of clustering accuracy and anomaly detection accuracy, while also showing great efficiency on a number of healthcare datasets as well as the Symptoms-to-Disease 7K dataset. DCMHTS, which combines trigger-based classification and deep clustering, helps provide a scalable, real-time, and resource-efficient framework for intelligent healthcare systems, indicating a step towards establishing more efficient disease detection and patient control.</p>