Deteksi Dini Hipertensi pada Usia Dewasa Menggunakan Deep Neural Network Berdasarkan Data Skrining Kesehatan

Q: What is the clinical evidence from this study?

Study design: Cross-Sectional. Population: Hypertension (n=214). Intervention: Deep Neural Network (DNN) classification model using health screening features vs. Actual hypertension status (defined as systolic BP >= 140 mmHg and/or diastolic BP >= 90 mmHg). Primary outcome: Testing dataset classification accuracy for early hypertension detection.

Q: What does this research mean for the field?

A deep neural network trained on basic community health screening indicators accurately identifies adult hypertension status, achieving 88.37% accuracy and a 93% F1-score. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

Sofiah Baiti Auliyah; Dian Ahkam Sani; Muhammad Udin

doi:10.31284/j.integer.2026.v11i2.9139

September 30, 2026INTEGER Journal of Information TechnologyOpen Access

Deteksi Dini Hipertensi pada Usia Dewasa Menggunakan Deep Neural Network Berdasarkan Data Skrining Kesehatan

View Full Paper

Ask AI

Bookmark

Share

Key result

Deep neural network detects adult hypertension with ~88% accuracy from routine community health screening data.

n=214

Why the study?

Hypertension is a leading cause of cardiovascular disease and mortality worldwide, making early detection essential to reduce complications and improve preventive healthcare. This study aimed to develop a Deep Neural Network model for early hypertension detection in adults using community health screening data.

Does a Deep Neural Network model accurately classify hypertension status based on community health screening data in adults?

Key Points

Develop and evaluate a Deep Neural Network model for early hypertension classification using community health screening records from adult individuals.
Analyzed health screening records from N=214 adults, utilizing seven input variables: gender, age, body weight, height, waist circumference, systolic blood pressure, and diastolic blood pressure.
Preprocessed data using label encoding, standardization via StandardScaler, and an 80% training to 20% testing split.
Built a deep architecture featuring three hidden layers with ReLU activation, a Sigmoid output layer, Adam optimizer, Binary Crossentropy loss, Dropout, and Early Stopping.
The proposed deep learning model achieved an overall accuracy of 88.37% on the testing dataset.
For the hypertension class, the model demonstrated a precision of 94%, recall of 91%, and an F1-score of 93%.

Study Design

Type

Cross-Sectional (n=214)

Multicenter

No

Structured PICO

Does a Deep Neural Network model accurately classify hypertension status based on community health screening data in adults?

P

Population

214 adult community health screening participants (aged 18-60 years, mean age 43.92 years) from Posyandu in Tebas Village, Pasuruan Regency, Indonesia.

E

Exposure

Deep Neural Network (DNN) classification model (3 hidden layers with 64, 32, and 16 neurons using ReLU activation, Sigmoid output layer, Dropout 0.2, Adam optimizer, Binary Crossentropy loss function, and Early Stopping) using 7 input features (gender, age, body weight, height, waist circumference, systolic blood pressure, and diastolic blood pressure).

O

Outcome

Classification performance for hypertension status (defined as systolic BP >=140 mmHg and/or diastolic BP >=90 mmHg) measured by accuracy, precision, recall, and F1-score on the testing dataset.surrogate

Limitations

Small overall sample size (214 records, with only 43 records in the testing set)
Data collected from a single village Posyandu, limiting geographic diversity
Class imbalance with fewer normal blood pressure records in the dataset, resulting in a lower F1-score (74%) for the normal class
Target variable (hypertension status) is directly derived from systolic and diastolic blood pressure, which were also included as input features
Smaller number of normal class records in the dataset, resulting in more limited learning information and lower performance (F1-score 74%) for the normal class compared to the hypertension class

Ask about this paper

Answers are grounded in this paper and cite their sources.

Implication

A Deep Neural Network model trained on basic community health screening variables achieved 88.37% accuracy in classifying hypertension status in Indonesian adults.

Abstract

Hypertension is one of the leading causes of cardiovascular disease and contributes significantly to mortality worldwide. Early detection is essential to reduce the risk of complications and improve preventive healthcare. This study aims to develop an early hypertension detection model for adults using the Deep Neural Network (DNN) algorithm based on community health screening data collected from Posyandu Desa Tebas, Gondangwetan District, Pasuruan Regency. The dataset consisted of 214 health records with seven input variables: gender, age, body weight, height, waist circumference, systolic blood pressure, and diastolic blood pressure. Data preprocessing included data cleaning, feature selection, label encoding, standardization using StandardScaler, and dataset splitting into 80% training data and 20% testing data. The proposed DNN architecture consisted of three hidden layers with ReLU activation, a Sigmoid output layer, Adam optimizer, Binary Crossentropy loss function, Dropout, and Early Stopping to reduce overfitting. Experimental results showed that the proposed model achieved an accuracy of 88.37%, with precision of 94%, recall of 91%, and F1-score of 93% for the hypertension class. These results indicate that the Deep Neural Network model can effectively classify hypertension status and has the potential to support early screening for hypertension.

Discussion

No takes yet. Share an insight, caveat, or question.

Cite This Study

Auliyah et al. (2026) conducted a cross-sectional in Hypertension (n=214). Deep Neural Network (DNN) classification model using health screening features vs. Actual hypertension status (defined as systolic BP >= 140 mmHg and/or diastolic BP >= 90 mmHg) was evaluated on Testing dataset classification accuracy for early hypertension detection. The Deep Neural Network model achieved 88.37% accuracy, 94% precision, 91% recall, and a 93% F1-score in classifying hypertension from adult community health screening data.

synapsesocial.com/papers/6ade998aa18383b786523774 https://doi.org/10.31284/j.integer.2026.v11i2.9139

View Full Paper

Ask AI

Bookmark

Share

Deteksi Dini Hipertensi pada Usia Dewasa Menggunakan Deep Neural Network Berdasarkan Data Skrining Kesehatan

Key Points

Study Design

Structured PICO

Limitations

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

Deteksi Dini Hipertensi pada Usia Dewasa Menggunakan Deep Neural Network Berdasarkan Data Skrining Kesehatan

Key Points

Study Design

Structured PICO

Limitations

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

Keep reading

Keep reading