CHRONICAL KIDNEY DISEASE PREVENTION

– USING MACHINE LEARNING

A PROJECT REPORT

Submitted by

MAHIMA(2021021144)
RUCHI RATHOR(2021021163)
of

Bachelor of Technology

in

Computer Science And Engineering

May 2024

Madan Mohan Malaviya University of Technology

GORAKHPUR - 273010

1|Page
 TABLE OF CONTENTS

S. No. Topic Page No

1. Abstract 3

2. Introduction 4

3. Literature Survey 5

4. Existing system 6

5. Proposed system 7

6. Prediction and Diet 8

7. Implementation 9

8. Stages of Chronic Kidney Disease 11

9. Result and Discussion 13

10. Conclusion 15

11. Reference 16

2|Page
 1. ABSTRACT

 In the modern world, chronic kidney disease has become one of the most
hazardous diseases. CKD is a condition in which the kidney cannot
perform the proper filtering of the blood or it stooped working completely
which causes the left toxic into the blood, which leads the patient to
death. It is likely impossible to detect CKD in the early stages, and it is
very difficult to save patient’s lives in the last stage of CKD. A patient's
life can be saved by renal transplant or the early detection the CKD.
Machine Learning algorithm techniques have played a very important
role in CKD prediction.
 Machine Learning (ML) Techniques like KNN, Decision Tree, and ANN
are used in this review. Generally, the majority of the algorithms are
based on supervised learning and classification problem solving.
 Their focus was not mainly on the specific stages prediction. In this
study, both binary and multi classification for stage prediction have been
carried out. The prediction models used include Random Forest (RF),
Support Vector Machine (SVM) and Decision Tree (DT).
 Analysis of variance and recursive feature elimination using cross
validation have been applied for feature selection. Evaluation of the
models was done using tenfold cross-validation. The results from the
experiments indicated that RF based on recursive feature elimination with
cross validation has better performance than SVM and DT.

3|Page
 2. INTRODUCTION

Chronic Kidney Disease (CKD) or chronic renal disease has become a major
issue with a steady growth rate. A person can only survive without kidneys for
an average time of 18 days, which makes a huge demand for a kidney transplant
and Dialysis. It is important to have effective methods for early prediction of
CKD. Machine learning methods are effective in CKD prediction. This work
proposes a workflow to predict CKD status based on clinical data, incorporating
data prepossessing, a missing value handling method with collaborative filtering
and attributes selection.
Out of the 11 machine learning methods considered, the extra tree classifier and
random forest classifier are shown to result in the highest accuracy and minimal
bias to the attributes. The research also considers the practical aspects of data
collection and highlights the importance of incorporating domain knowledge
when using machine learning for CKD status prediction.
National kidney foundation classifies stages of CKD into five based on the
abnormal kidney function and reduced Glomerular Filtration Rate (GFR), which
measures a level of kidney function,. The mildest stage (stage 1 and stage 2) is
known with only a few symptoms and stage 5 is considered as end-stage or
kidney failure. The Renal Replacement Therapy (RRT) cost for total kidney
failure is very expensive.
The treatment is not also available in most developing countries like Ethiopia.
As a result, the management of kidney failure and its complications is very
difficult in developing countries due to shortage of facilities, physicians, and the
high cost to get the treatment [4, 5]. Hence, early detection of CKD is very
essential to minimize the economic burden and maximize the effectiveness of
treatments.

4|Page
 3. Literature Survey

Chronic Kidney Disease (CKD) represents a global health challenge, with a

profound impact on patient morbidity, mortality, and healthcare systems
worldwide. A comprehensive literature survey reveals a diverse array of
research focusing on various aspects of CKD, including epidemiology, risk
factors, pathophysiology, clinical manifestations, diagnosis, treatment, and
preventive strategies.

Risk Factors :
Research elucidates the multifactorial etiology of CKD, with risk factors
including diabetes mellitus, hypertension, obesity, smoking, genetic
predisposition, and environmental factors. Studies explore the interplay
between these risk factors and underlying pathophysiological mechanisms,
contributing to the progression of kidney dysfunction.

Pathophysiological Mechanisms:
Investigations into the pathophysiology of CKD uncover complex
mechanisms involving hemodynamic alterations, inflammation, oxidative
stress, fibrosis, and endothelial dysfunction. Molecular and cellular studies
provide insights into the interplay between renal injury pathways, offering
potential targets for therapeutic interventions.

Preventive Measures and Public Health Interventions:

Public health initiatives focus on primary prevention strategies to reduce the
incidence and progression of CKD, emphasizing the importance of early
detection, risk factor modification, health education, and screening
programs. Population-based studies evaluate the effectiveness of
interventions in reducing CKD burden and improving outcomes at the
community level.

5|Page
 [Link] System:

The existing system for preventing Chronic Kidney Disease (CKD) involves a
combination of population-based strategies, targeted interventions for high-risk
individuals, and public health initiatives aimed at reducing modifiable risk
factors. Here's an overview of the existing CKD prevention system
Public health campaigns raise awareness about CKD risk factors, symptoms,
and preventive measures through mass media, community outreach programs,
and educational materials.
These campaigns aim to educate the general population about lifestyle
modifications, such as maintaining a healthy diet, regular exercise, smoking
cessation, and limiting alcohol consumption, to reduce the risk of developing
CKD.
Healthcare systems implement screening programs to identify individuals at risk
for CKD, including those with diabetes, hypertension, obesity, and a family
history of kidney disease.
Routine health screenings, such as blood pressure measurements, urine tests for
proteinuria, and serum creatinine measurements, enable early detection of CKD
and prompt initiation of preventive measures.

Limitations of Existing System:

 Late Diagnosis and Underdiagnosis
 Limited Access to Care:
 Treatment Adherence Challenges
 High Treatment Costs
 Limited Treatment Options

6|Page
 5. Proposed System:

Proposed System: Enhanced Chronic Kidney Disease (CKD) Prevention

Framework
The proposed system aims to build upon the existing CKD prevention strategies
by incorporating innovative approaches, leveraging digital health technologies,
and fostering collaboration among stakeholders. Here's an outline of the
proposed CKD prevention framework:
Implement community-based CKD awareness campaigns targeting high-risk
populations, ethnic minorities, and underserved communities.

7|Page
 Fig1: Architecture of proposed system in kidney disease
6. Chronic kidney disease prediction and diet
recommendation using machine learning:

Several machine learning techniques can be employed in the Chronic Kidney

Disease (CKD) prevention system to predict, classify, and identify individuals
at risk. Here are some commonly used techniques:
 Logistic Regression: Logistic regression is a statistical method used for
binary classification tasks, such as predicting the presence or absence of
CKD based on patient demographics, clinical measurements, and
laboratory test results.
 Decision Trees: Decision trees are intuitive and interpretable machine
learning models that recursively split the data into subsets based on the
most informative features, allowing for the classification of CKD risk
based on hierarchical decision rules.
 Random Forests: Random forests are ensemble learning methods that
combine multiple decision trees to improve predictive accuracy and
robustness. They can handle non-linear relationships and interactions
between variables, making them suitable for CKD risk prediction tasks.
 Support Vector Machines (SVM): SVM is a supervised learning
algorithm used for classification tasks that aims to find the hyperplane
that best separates classes in high-dimensional feature space. SVMs are
effective for CKD risk prediction when dealing with complex, non-linear
decision boundaries.
 Gradient Boosting Machines (GBM): GBM is an ensemble learning
technique that builds a sequence of weak learners (typically decision
trees) to iteratively improve model performance. GBM can capture
complex relationships and interactions in the data, making it suitable for
CKD risk prediction tasks.
 K-Nearest Neighbors (KNN): KNN is a simple and intuitive
classification algorithm that assigns a class label to an input sample based
on the majority class among its k nearest neighbors in the feature space.
KNN can be used for CKD risk prediction based on similarity to known
cases.

8|Page
  Naive Bayes Classifier: Naive Bayes is a probabilistic classification
algorithm based on Bayes' theorem and the assumption of conditional
independence between features. Despite its simplicity, Naive Bayes can
provide competitive performance for CKD risk prediction tasks.

7. Implementation

Implementing a Chronic Kidney Disease (CKD) prevention system involves

several steps, including data collection, preprocessing, model development,
evaluation, deployment, and monitoring. Here's an outline of the
implementation process:

 Data Collection and Preprocessing: Collect diverse datasets containing

demographic information, clinical variables (e.g., blood pressure, serum
creatinine), genetic markers, lifestyle factors, and environmental
exposures relevant to CKD risk.

 Model Development: Choose appropriate machine learning algorithms

(e.g., logistic regression, decision trees, random forests, support vector
machines) based on the nature of the CKD prediction task and the
characteristics of the dataset.

 Model Evaluation: Evaluate the trained models using appropriate

evaluation metrics (e.g., accuracy, precision, recall, F1-score, area under
the receiver operating characteristic curve) on the validation set.

 Model Deployment: Deploy the trained machine learning models into

production environments, such as healthcare IT systems, clinical decision
support systems, or mobile health applications.

 User Training and Support: Provide training and support to end-users,

Including healthcare providers, patients, and system administrators, on
how to use the CKD prevention system effectively. Develop user-friendly
interfaces, documentation, and tutorials to facilitate adoption and
usability of the system.

Integrate the models with existing healthcare workflows and data infrastructure,
ensuring compatibility with electronic health records (EHRs) and
interoperability standards.
9|Page
 Fig2: Machine Learning Process Flow

Fig3: Causes of Chronic Kidney Disease

10 | P a g e
 Fig4: Features of CKD

[Link] of Chronic Kidney Disease:

It's important to note that CKD progression can vary widely among individuals,
and early detection and management are crucial for slowing disease
progression, preventing complications, and preserving kidney function. Regular
monitoring, lifestyle modifications, medication management, and timely referral
to nephrology care are essential components of CKD management at every
stage.

Stage 1: Kidney Damage with Normal or Increased GFR (≥90 mL/min/1.73

m²):

Stage 2: Kidney Damage with Mildly Decreased GFR (60-89 mL/min/1.73 m²):

Stage 3: Moderate Decrease in GFR (30-59 mL/min/1.73 m²):

Stage 4: Severe Decrease in GFR (15-29 mL/min/1.73 m²):

11 | P a g e
Stage 5: Kidney Failure (GFR <15 mL/min/1.73 m²):

12 | P a g e
 9. Results & Discussion:

The proposed innovative solutions for CKD management hold

significant promise for revolutionizing the care paradigm and
improving patient outcomes. Here, we discuss the potential results
and implications of implementing these strategies:

1. Personalized Risk Assessment and Prevention:

- Result: Early identification of individuals at high risk for CKD
development or progression allows for targeted interventions to
mitigate risk factors and delay disease onset or progression.
- Discussion: By leveraging AI algorithms and predictive analytics,
healthcare providers can identify at-risk individuals more effectively
and tailor preventive strategies to address individualized risk factors.
This proactive approach has the potential to reduce the incidence and
burden of CKD on healthcare systems.

2. Remote Monitoring and Telehealth Solutions:

- Result: Remote monitoring platforms and telehealth solutions
enable real-time tracking of key CKD parameters, facilitating
proactive intervention and timely adjustments to treatment plans.
- Discussion: Remote monitoring enhances patient engagement and
adherence to care plans by providing continuous support and
feedback. Telehealth consultations improve access to specialized care,
particularly for individuals in rural or underserved areas, thereby
reducing barriers to optimal CKD management.

13 | P a g e
3. Precision Medicine and Biomarker-Based Monitoring:
- Result: Precision medicine approaches, including genomic
profiling and biomarker analysis, enable personalized risk
stratification, treatment selection, and monitoring of CKD
progression.
- Discussion: Biomarker-based monitoring enhances the precision
and timeliness of CKD diagnosis and progression assessment,
allowing for earlier intervention and targeted therapies. This approach
optimizes resource allocation and improves clinical outcomes by
tailoring treatment to individual patient needs.

4. Collaborative Care Coordination and Decision Support:

- Result: Multidisciplinary care teams and clinical decision support
systems facilitate coordinated care, evidence-based decision-making,
and seamless communication among healthcare providers.
- Discussion: Collaborative care models enhance care coordination,
optimize treatment outcomes, and improve patient satisfaction by
leveraging the expertise of diverse healthcare professionals.

In conclusion, the proposed innovative solutions for CKD

management offer a transformative approach to care delivery,
emphasizing personalized interventions, remote monitoring,
collaborative care coordination, and patient empowerment. By
embracing these advancements, healthcare systems can enhance the
effectiveness, efficiency, and patient-centeredness of CKD
management, ultimately improving outcomes and reducing the burden
of this chronic condition on individuals and healthcare systems alike.

14 | P a g e
 10. Conclusion

Early prediction is very crucial for both experts and patients to prevent and slow
down the progress of chronic kidney disease to kidney failure. Applying the
models on the original dataset, we have got the highest accuracy with RF, SVM,
and XGBoost. The accuracy was 99.8% for the binary class and 82.56% for
five-class. DT produced lowest performance compared to RF. RF also produced
the highest f1_score values. SVM and RF with RFECV produced the highest
accuracy of 99.8%for binary class. XGBoost has 82.56% accuracy for five-class
datasets which is the highest. Hence we believe that multi classification work
was very important to know the stages of the disease and suggest needed
treatments for the patients in order to save their lives.

Precision medicine approaches, including genomic profiling and biomarker-

based monitoring, offer tailored treatment regimens and enable proactive
management of CKD progression. Collaborative care coordination, facilitated
by multidisciplinary teams and decision support systems, ensures
comprehensive and evidence-based care delivery, optimizing treatment
outcomes and patient satisfaction.

In essence, by embracing these innovative solutions and fostering

multidisciplinary collaboration, healthcare systems can transform the CKD
management landscape, enhance patient outcomes, and alleviate the burden of
this chronic condition on individuals, families, and society as a whole. Through
continued research, implementation, and adaptation, we can strive towards a
future where CKD is effectively prevented, managed, and ultimately,
conquered.

15 | P a g e
 11. References

 Data Preparation for Statistical Modeling and Machine

Learning. [Link]
tutorials/data-preparation-for-machine-learning/.
 Oliver T. Machine Learning For Absolute Beginners.
 Radhakrishnan J, Mohan S. KI Reports and World Kidney Day.
Kidney Int Reports.
 Ethiopia: kidney
disease. [Link]
disease.
 Agrawal A, Agrawal H, Mittal S, Sharma M. Disease Prediction
Using Machine Learning. SSRN Electron J.
 ZarPhyu T, Oo NN. Performance comparison of feature selection
methods. MATEC Web Conf. 2016;42
 Vidhya A. Introduction to Feature Selection methods with an example
(or how to select the right
variables?). [Link]
on-to-feature-selection-methods-with-an-example-or-how-to-select-
the-right-variables/.

16 | P a g e