Araştırma Makalesi
BibTex RIS Kaynak Göster

Comparison of Combining Methods for Multiple Continuous Diagnostic Tests Using ROC Curve Analysis

Yıl 2020, , 553 - 567, 15.09.2020
https://doi.org/10.20515/otd.673372

Öz

In the field of medicine, despite the
uncertainty about incomplete clinical information and clinical outcomes, it is
necessary to make accurate and logical decisions about the treatment and care
of the patient. Medical tests have an important role in terms of medical
diagnosis, but they also provide indispensable benefits in terms of planning of
correct treatment and reduction of treatment costs. Providing reliable
information about the patient's condition, correctly classifying patients and
healthy units, and providing positive contributions to healthcare staff's
correct planning for the treatment of the patient constitute the main
objectives of the medical test. One of the widely used statistical techniques
for evaluating the performance of the classification rule is the Receiver
Operating Characteristic (ROC) curve. Nowadays, rather than using a single
diagnostic test result to determine the diagnosis of the disease, it is
possible to make more precise diagnosis or classification using more than one
test. Different diagnostic tests in health field studies are sensitive to
different aspects of the disease. Therefore, a single diagnostic test can not
always be relied upon to assess an individual's health condition. It is
possible to make more accurate and closer classification using more than one
diagnostic test. Various models have been proposed for this purpose. These
models are best linear combination method, linear discriminant analysis,
quadratic discriminant analysis and logistic discriminant analysis methods. The
purpose of this study is to demonstrate the utility of the logistic regression
model as an alternative to the methods used in the combination of diagnostic
tests in combining diagnostic tests of linear and quadratic discriminant
analysis. The second objective is to compare the performance of combining
diagnostic tests of logistic, linear and quadratic discriminant analyzes with
parametric and nonparametric ROC analyzes of the performance of the best linear
combination method with the minimax procedure. It is to determine which method
performs well in cases where the assumption of the multivariate normal
distribution is realized and not realized.

Kaynakça

  • Nicoll, D., McPhee, S. J., & Pignone, M. (2012). Pocket guide to diagnostic tests. McGraw-Hill Medical.
  • Epstein, A. M., Begg, C. B., & McNeil, B. J. (1986). The use of ambulatory testing in prepaid and fee-for-service group practices. New England Journal of Medicine, 314(17), 1089-1094.
  • Zhang, D. D., Zhou, X. H., Freeman, D. H., & Freeman, J. L. (2002). A non‐parametric method for the comparison of partial areas under ROC curves and its application to large health care data sets. Statistics in medicine, 21(5), 701-715.
  • Friedman, J., Hastie, T., & Tibshirani, R. (2001). The elements of statistical learning (Vol. 1): Springer series in statistics New York.
  • Colak, E., Mutlu, F., Bal, C., Oner, S., Ozdamar, K., Gok, B., & Cavusoglu, Y. (2012). Comparison of semiparametric, parametric, and nonparametric ROC analysis for continuous diagnostic tests using a simulation study and acute coronary syndrome data. Computational and mathematical methods in medicine, 2012.
  • Hanley, J. A., & McNeil, B. J. (1982). The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology, 143(1), 29-36.
  • Metz, C. E. (1978). Basic principles of ROC analysis. Paper presented at the Seminars in nuclear medicine.
  • Fawcett, T. (2006). An introduction to ROC analysis. Pattern Recognition Letters, 27(8), 861-874. doi:http://dx.doi.org/10.1016/ j.patrec.2005.10.010
  • Li, J., & Fine, J. P. (2008). ROC analysis with multiple classes and multiple tests: methodology and its application in microarray studies. Biostatistics, 9(3), 566-576. doi:10.1093/biostatistics/kxm050
  • Su, J. Q., & Liu, J. S. (1993). Linear-Combinations of Multiple Diagnostic Markers. Journal of the American Statistical Association, 88(424), 1350-1355.
  • Reiser, B., & Faraggi, D. (1997). Confidence intervals for the generalized ROC criterion. Biometrics, 644-652.
  • Pepe, M. S., & Thompson, M. L. (2000). Combining diagnostic test results to increase accuracy. Biostatistics, 1(2), 123-140.
  • Liu, C., Liu, A., & Halabi, S. (2011). A min–max combination of biomarkers to improve diagnostic accuracy. Statistics in Medicine, 30(16), 2005-2014.
  • Gao, F., Xiong, C., Yan, Y., Yu, K., & Zhang, Z. (2008). Estimating optimum linear combination of multiple correlated diagnostic tests at a fixed specificity with receiver operating characteristic curves. Journal of Data Science, 6(1), 105-123.
  • Yu, W., Kwon, M. S., & Park, T. (2015). Multivariate quantitative multifactor dimensionality reduction for detecting gene-gene interactions. Human heredity, 79(3-4), 168-181.
  • Kang, L., Liu, A., & Tian, L. (2016). Linear combination methods to improve diagnostic/prognostic accuracy on future observations. Stat Methods Med Res, 25(4), 1359-1380. doi:10.1177/0962280213481053
  • Sameera, G., Vardhan, R. V., & Sarma, K. V. S. (2016). Binary classification using multivariate receiver operating characteristic curve for continuous data. Journal of biopharmaceutical statistics, 26(3), 421-431.
  • Mamtani, M. R., Thakre, T. P., Kalkonde, M. Y., Amin, M. A., Kalkonde, Y. V., Amin, A. P., & Kulkarni, H. (2006). A simple method to combine multiple molecular biomarkers for dichotomous diagnostic classification. BMC bioinformatics, 7(1), 442.
  • Ma, S., & Huang, J. (2007). Combining multiple markers for classification using ROC. Biometrics, 63(3), 751-757. doi:10.1111/j.1541-0420.2006.00731.x
  • Qin, J., & Zhang, B. (2010). Best combination of multiple diagnostic tests for screening purposes. Statistics in Medicine, 29(28), 2905-2919.
  • Lin, H., Zhou, L., Peng, H., & Zhou, X. H. (2011). Selection and combination of biomarkers using ROC method for disease classification and prediction. Canadian Journal of Statistics, 39(2), 324-343.
  • Johnson, R. A., & Wichern, D. (2002). Multivariate analysis. John Wiley & Sons, Ltd.
  • Schulzer, M. (1994). Diagnostic tests: a statistical review. Muscle & nerve, 17(7), 815-819.
  • Segen, J. C., & Wade, J. (2002). The patient's guide to medical tests: everything you need to know about the tests your doctor orders. Infobase Publishing.
  • Anderson, T. W., & Bahadur, R. R. (1962). Classification into two multivariate normal distributions with different covariance matrices. The annals of mathematical statistics, 33(2), 420-431.

Sürekli Yapıdaki Çoklu Tanı Testleri İçin Birleştirme Yöntemlerinin ROC Eğrisi Analizi Kullanarak Karşılaştırılması

Yıl 2020, , 553 - 567, 15.09.2020
https://doi.org/10.20515/otd.673372

Öz

Tıp alanında, eksik klinik
bilgilere ve klinik sonuçlarla ilgili belirsizliğe rağmen, hastanın tedavisi ve
bakımı ile ilgili doğru ve mantıklı kararlar vermek gerekmektedir. Medikal
testler tıbbi tanı açısından önemli bir role sahip olmanın yanı sıra doğru
tedavinin planlanması ile tedavi maliyetlerinin azaltılması yönünden göz ardı
edilemeyecek katkılar sağlamaktadırlar. Hastanın durumu hakkında güvenilir
bilgi sağlamak, hasta ile sağlıklı birimleri doğru sınıflamak ve hastanın
tedavisi için sağlık personelinin doğru planlama yapmasına olumlu katkılar
sağlamak medikal testin temel amaçlarını oluşturmaktadır. Sınıflandırma
kuralının performansını değerlendirmek için yaygın olarak kullanılan
istatistiksel tekniklerden biri, Receiver Operating Characteristic (ROC)
eğrisidir. Günümüzde, tek bir tanı testi sonucu kullanılarak hastalığa ait
tanının belirlenmesi yerine birden fazla test kullanılarak daha kesin bir tanı
ya da sınıflandırma yapmak mümkündür. Sağlık alanında yapılan çalışmalarda
farklı tanı testleri, hastalığın farklı yönlerine duyarlıdır. Dolayısıyla,
bireyin sağlık durumunu değerlendirmek için her zaman tek bir tanı testine
güvenilemez. Birden fazla tanı testi kullanılarak gerçeğe daha yakın ve doğru
sınıflandırma yapmak mümkündür. Bu amaç doğrultusunda çeşitli modeller öne
sürülmüştür. Bu modeller, en iyi doğrusal birleştirme yöntemi, doğrusal ayırma
analizi, karesel ayırma analizi ve lojistik ayırma analizi yöntemleridir. Bu
çalışmasının amacı, tanı testlerinin birleştirilmesinde kullanılan yöntemlere
alternatif olarak lojistik regresyon modeli ile doğrusal ve karesel ayırma
analizlerinin tanı testlerinin birleştirilmesinde kullanılabilirliğini
göstermek, lojistik, doğrusal ve karesel ayırma analizlerinin tanı testlerinin
birleştirilmesindeki performanslarını, minimax prosedürü ile en iyi doğrusal
birleştirme yönteminin performansını parametrik ve parametrik olmayan ROC
analizleri ile karşılaştırmak, çok değişkenli normal dağılım varsayımının
gerçekleştiği ve gerçekleşmediği durumlarda hangi yöntemin iyi performans
gösterdiğini belirlemektir.

Kaynakça

  • Nicoll, D., McPhee, S. J., & Pignone, M. (2012). Pocket guide to diagnostic tests. McGraw-Hill Medical.
  • Epstein, A. M., Begg, C. B., & McNeil, B. J. (1986). The use of ambulatory testing in prepaid and fee-for-service group practices. New England Journal of Medicine, 314(17), 1089-1094.
  • Zhang, D. D., Zhou, X. H., Freeman, D. H., & Freeman, J. L. (2002). A non‐parametric method for the comparison of partial areas under ROC curves and its application to large health care data sets. Statistics in medicine, 21(5), 701-715.
  • Friedman, J., Hastie, T., & Tibshirani, R. (2001). The elements of statistical learning (Vol. 1): Springer series in statistics New York.
  • Colak, E., Mutlu, F., Bal, C., Oner, S., Ozdamar, K., Gok, B., & Cavusoglu, Y. (2012). Comparison of semiparametric, parametric, and nonparametric ROC analysis for continuous diagnostic tests using a simulation study and acute coronary syndrome data. Computational and mathematical methods in medicine, 2012.
  • Hanley, J. A., & McNeil, B. J. (1982). The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology, 143(1), 29-36.
  • Metz, C. E. (1978). Basic principles of ROC analysis. Paper presented at the Seminars in nuclear medicine.
  • Fawcett, T. (2006). An introduction to ROC analysis. Pattern Recognition Letters, 27(8), 861-874. doi:http://dx.doi.org/10.1016/ j.patrec.2005.10.010
  • Li, J., & Fine, J. P. (2008). ROC analysis with multiple classes and multiple tests: methodology and its application in microarray studies. Biostatistics, 9(3), 566-576. doi:10.1093/biostatistics/kxm050
  • Su, J. Q., & Liu, J. S. (1993). Linear-Combinations of Multiple Diagnostic Markers. Journal of the American Statistical Association, 88(424), 1350-1355.
  • Reiser, B., & Faraggi, D. (1997). Confidence intervals for the generalized ROC criterion. Biometrics, 644-652.
  • Pepe, M. S., & Thompson, M. L. (2000). Combining diagnostic test results to increase accuracy. Biostatistics, 1(2), 123-140.
  • Liu, C., Liu, A., & Halabi, S. (2011). A min–max combination of biomarkers to improve diagnostic accuracy. Statistics in Medicine, 30(16), 2005-2014.
  • Gao, F., Xiong, C., Yan, Y., Yu, K., & Zhang, Z. (2008). Estimating optimum linear combination of multiple correlated diagnostic tests at a fixed specificity with receiver operating characteristic curves. Journal of Data Science, 6(1), 105-123.
  • Yu, W., Kwon, M. S., & Park, T. (2015). Multivariate quantitative multifactor dimensionality reduction for detecting gene-gene interactions. Human heredity, 79(3-4), 168-181.
  • Kang, L., Liu, A., & Tian, L. (2016). Linear combination methods to improve diagnostic/prognostic accuracy on future observations. Stat Methods Med Res, 25(4), 1359-1380. doi:10.1177/0962280213481053
  • Sameera, G., Vardhan, R. V., & Sarma, K. V. S. (2016). Binary classification using multivariate receiver operating characteristic curve for continuous data. Journal of biopharmaceutical statistics, 26(3), 421-431.
  • Mamtani, M. R., Thakre, T. P., Kalkonde, M. Y., Amin, M. A., Kalkonde, Y. V., Amin, A. P., & Kulkarni, H. (2006). A simple method to combine multiple molecular biomarkers for dichotomous diagnostic classification. BMC bioinformatics, 7(1), 442.
  • Ma, S., & Huang, J. (2007). Combining multiple markers for classification using ROC. Biometrics, 63(3), 751-757. doi:10.1111/j.1541-0420.2006.00731.x
  • Qin, J., & Zhang, B. (2010). Best combination of multiple diagnostic tests for screening purposes. Statistics in Medicine, 29(28), 2905-2919.
  • Lin, H., Zhou, L., Peng, H., & Zhou, X. H. (2011). Selection and combination of biomarkers using ROC method for disease classification and prediction. Canadian Journal of Statistics, 39(2), 324-343.
  • Johnson, R. A., & Wichern, D. (2002). Multivariate analysis. John Wiley & Sons, Ltd.
  • Schulzer, M. (1994). Diagnostic tests: a statistical review. Muscle & nerve, 17(7), 815-819.
  • Segen, J. C., & Wade, J. (2002). The patient's guide to medical tests: everything you need to know about the tests your doctor orders. Infobase Publishing.
  • Anderson, T. W., & Bahadur, R. R. (1962). Classification into two multivariate normal distributions with different covariance matrices. The annals of mathematical statistics, 33(2), 420-431.
Toplam 25 adet kaynakça vardır.

Ayrıntılar

Birincil Dil Türkçe
Konular Sağlık Kurumları Yönetimi
Bölüm ORİJİNAL MAKALELER / ORIGINAL ARTICLES
Yazarlar

Muzaffer Bilgin 0000-0002-6072-6466

Aşkın Doğan 0000-0001-5334-6265

Ertugrul Colak 0000-0003-3251-1043

Yayımlanma Tarihi 15 Eylül 2020
Yayımlandığı Sayı Yıl 2020

Kaynak Göster

Vancouver Bilgin M, Doğan A, Colak E. Sürekli Yapıdaki Çoklu Tanı Testleri İçin Birleştirme Yöntemlerinin ROC Eğrisi Analizi Kullanarak Karşılaştırılması. Osmangazi Tıp Dergisi. 2020;42(5):553-67.


13299        13308       13306       13305    13307  1330126978