Estimation of Risk Factors Associated with Breast Cancer Worry with Using Artificial Neural Network Model

Gülçin Nacar; Feyza İnceoğlu; Sermin Timur

Article Details

Estimation of Risk Factors Associated with Breast Cancer Worry with Using Artificial Neural Network Model

Authors:

Gülçin Nacar | Feyza İnceoğlu | Sermin Timur

Publisher:

Ordu University

Main Topics:

Hemşirelik

Journal:

Ordu Üniversitesi Hemşirelik Çalışmaları Dergisi Vol.6 Issue 2, pp.253-264

Document Type:

Article

Publication Date:

1/2023

Document Language:

Turkish

Document ID (Org No.):

2001297934

pISSN:

2667-6052

Keywords:

Abstract:

 Objective: In this study, it was aimed to determine the risk factors associated with breast cancer worry in women using binary logistic regression and artificial neural network ANN models. Method: The study was conducted with 365 women aged 18 and over, reached in a Family Health Center. Accuracy rate and the area under the ROC curve were used to determine the performance of the multilayer perceptron neural network model used to identify the factors associated with breast cancer worry. Research data were collected using a personal information form and Breast Cancer Worry Scale BCWC. Results: In the univariate sample tests, it was found that the BCWS scores showed statistically significant differences according to the variables of age, income, menopause and smoking p0.05. The correct classification rates for breast cancer worry of the created multilayer perceptual neural network model were calculated as 90.9 in the training dataset and 89 in the test dataset. Considering the importance values of the variables; educational status 98.9 was found to be the most influential factor on breast cancer worry. In the binary logistic regression analysis, it was found that income status had a 2.384 fold effect on breast cancer worry OR 2.384, 95 CI 1.010-5.628. It is recommended that health professionals consider the identified risk factors when evaluating womens breast cancer worries. Conclusion: In the Binary Logistic Regression model, the income status was found to be effective on the BCWC as in the univariate sample tests. Educational status, parity, and occupation variables, which are the highest risk factors in ANN analysis, were not found significant in univariate statistical tests. It has been determined that ANN analyses prevent data loss in parametric tests. It is recommended that health professionals consider the identified risk factors when evaluating womens breast cancer worries.

 Objective: In this study, it was aimed to determine the risk factors associated with breast cancer worry in women using binary logistic regression and artificial neural network ANN models. Method: The study was conducted with 365 women aged 18 and over, reached in a Family Health Center. Accuracy rate and the area under the ROC curve were used to determine the performance of the multilayer perceptron neural network model used to identify the factors associated with breast cancer worry. Research data were collected using a personal information form and Breast Cancer Worry Scale BCWC. Results: In the univariate sample tests, it was found that the BCWS scores showed statistically significant differences according to the variables of age, income, menopause and smoking p0.05. The correct classification rates for breast cancer worry of the created multilayer perceptual neural network model were calculated as 90.9 in the training dataset and 89 in the test dataset. Considering the importance values of the variables; educational status 98.9 was found to be the most influential factor on breast cancer worry. In the binary logistic regression analysis, it was found that income status had a 2.384 fold effect on breast cancer worry OR 2.384, 95 CI 1.010-5.628. It is recommended that health professionals consider the identified risk factors when evaluating womens breast cancer worries. Conclusion: In the Binary Logistic Regression model, the income status was found to be effective on the BCWC as in the univariate sample tests. Educational status, parity, and occupation variables, which are the highest risk factors in ANN analysis, were not found significant in univariate statistical tests. It has been determined that ANN analyses prevent data loss in parametric tests. It is recommended that health professionals consider the identified risk factors when evaluating womens breast cancer worries.

 Objective: In this study, it was aimed to determine the risk factors associated with breast cancer worry in women using binary logistic regression and artificial neural network ANN models. Method: The study was conducted with 365 women aged 18 and over, reached in a Family Health Center. Accuracy rate and the area under the ROC curve were used to determine the performance of the multilayer perceptron neural network model used to identify the factors associated with breast cancer worry. Research data were collected using a personal information form and Breast Cancer Worry Scale BCWC. Results: In the univariate sample tests, it was found that the BCWS scores showed statistically significant differences according to the variables of age, income, menopause and smoking p0.05. The correct classification rates for breast cancer worry of the created multilayer perceptual neural network model were calculated as 90.9 in the training dataset and 89 in the test dataset. Considering the importance values of the variables; educational status 98.9 was found to be the most influential factor on breast cancer worry. In the binary logistic regression analysis, it was found that income status had a 2.384 fold effect on breast cancer worry OR 2.384, 95 CI 1.010-5.628. It is recommended that health professionals consider the identified risk factors when evaluating womens breast cancer worries. Conclusion: In the Binary Logistic Regression model, the income status was found to be effective on the BCWC as in the univariate sample tests. Educational status, parity, and occupation variables, which are the highest risk factors in ANN analysis, were not found significant in univariate statistical tests. It has been determined that ANN analyses prevent data loss in parametric tests. It is recommended that health professionals consider the identified risk factors when evaluating womens breast cancer worries.

 Amaç: Bu çalışmada, kadınlarda meme kanseri endişesi ile ilişkili risk faktörlerinin binary lojistik regresyon ve yapay sinir ağı YSA modelleri kullanılarak belirlenmesi amaçlanmıştır. Yöntem: Araştırma, bir aile sağlığı merkezinde, 18 yaş ve üzerinde olan 365 kadın ile 30 Nisan-15 Haziran 2021 tarihleri arasında yapılmıştır. Meme kanseri endişesi ile ilişkili faktörleri belirlemek için kullanılan çok katmanlı algılayıcı yapay sinir ağı modelinin performansını belirlemede, doğruluk oranı ve ROC eğrisinin altındaki alan kullanılmıştır. Araştırma verileri, Kişisel Bilgi Formu ve Meme Kanseri Endişe Skalası MKES kullanılarak toplanmıştır. Bulgular: Tek değişkenli örneklem testlerinde MKES puanlarının yaş, gelir durumu, menopoz ve sigara içme değişkenlerine göre istatistiksel olarak anlamlı farklılık gösterdiği bulunmuştur p0.05. Oluşturulan çok katmanlı algısal sinir ağı modelinin meme kanseri endişesi için doğru sınıflandırma oranları eğitim veri setinde  90.9 ve test veri setinde  89 olarak hesaplanmıştır. Değişkenlerin önem değerleri dikkate alındığında; meme kanseri endişesi üzerinde en yüksek düzeyde etkili faktörün eğitim durumu 98.9 olduğu bulunmuştur. Binary lojistik regresyon analizinde ise gelir durumunun meme kanseri endişesi üzerinde 2.384 kat etkili olduğu bulunmuştur OR 2.384, 95 CI 1.010-5.628. Sonuç: Binary Lojistik Regresyon modelinde gelir durumu tek değişkenli örneklem testlerinde olduğu gibi MKES üzerinde etkili bulunmuştur. YSA analizinde en yüksek risk faktörü olan eğitim durumu, parite ve meslek değişkenleri tek değişkenli istatistiksel testlerde anlamlı bulunmamıştır. YSA analizlerinin parametrik testlerde var olan veri kayıplarını önlediği saptanmıştır. Sağlık profesyonellerinin kadınların meme kanseri endişesini değerlendirirken saptanan risk faktörlerini göz önünde bulundurmaları önerilmektedir.

Article Details

Estimation of Risk Factors Associated with Breast Cancer Worry with Using Artificial Neural Network Model

Authors:

Gülçin Nacar | Feyza İnceoğlu | Sermin Timur

Publisher:

Ordu University

Main Topics:

Hemşirelik

Journal:

Ordu Üniversitesi Hemşirelik Çalışmaları Dergisi Vol.6 Issue 2, pp.253-264

Document Type:

Article

Publication Date:

1/2023

Document Language:

Turkish

Document ID (Org No.):

2001297934

pISSN:

2667-6052

Breast cancer worry

artificial neural network model

associated factorsMeme kanseri endişesi

risk faktörleri

yapay sinir ağları

Issues

Article Details

Estimation of Risk Factors Associated with Breast Cancer Worry with Using Artificial Neural Network Model

Authors: Gülçin Nacar | Feyza İnceoğlu | Sermin Timur

Publisher: Ordu University

Main Topics: Hemşirelik

Journal: Ordu Üniversitesi Hemşirelik Çalışmaları Dergisi Vol.6 Issue 2, pp.253-264

Document Type: Article

Publication Date: 1/2023

Document Language: Turkish

Document ID (Org No.): 2001297934

pISSN: 2667-6052

Breast cancer worry

artificial neural network model

associated factorsMeme kanseri endişesi

risk faktörleri

yapay sinir ağları

Issues

Authors:

Gülçin Nacar | Feyza İnceoğlu | Sermin Timur

Publisher:

Ordu University

Main Topics:

Hemşirelik

Journal:

Ordu Üniversitesi Hemşirelik Çalışmaları Dergisi Vol.6 Issue 2, pp.253-264

Document Type:

Article

Publication Date:

1/2023

Document Language:

Turkish

Document ID (Org No.):

2001297934

pISSN:

2667-6052