Pemodelan Recurrent Event Survival Dengan Prentice-William-Peterson Untuk Kejadian Stroke Berulang Di RSUD Haji Provinsi Jawa Timur

Royyanah, Atika Nur (2026) Pemodelan Recurrent Event Survival Dengan Prentice-William-Peterson Untuk Kejadian Stroke Berulang Di RSUD Haji Provinsi Jawa Timur. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003241033-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Analisis survival lazim digunakan untuk mempelajari waktu terjadinya peristiwa klinis. Pada banyak studi medis, kejadian yang sama dapat berulang pada individu yang sama sehingga dinamika risikonya bergantung pada urutan dan jarak antarkejadian. Penelitian ini menerapkan pemodelan kejadian berulang (recurrent events) menggunakan dua kerangka skala waktu, yaitu Prentice-Williams-Peterson Total Time (PWP-TT) dan Gap Time (PWP-GT). Penaksiran parameter dilakukan dengan memaksimumkan fungsi partial likelihood terstratifikasi melalui iterasi numerik Newton-Raphson. Perbedaan mendasar kedua pendekatan terletak pada definisi risk set di tiap strata kejadian. Hasil model menunjukkan bahwa model Extended Cox PWP-TT terpilih sebagai model terbaik berdasarkan nilai Akaike Information Criterion (AIC) terkecil dibandingkan model PWP-GT. Interpretasi model terpilih difokuskan pada urutan kejadian pertama dan kedua, sedangkan urutan kejadian ketiga tidak disertakan dalam pembahasan substantif karena keterbatasan ukuran sampel (n = 11) menyebabkan estimasi parameter pada strata tersebut tidak stabil secara numerik. Hasil model Extended Cox PWP-TT terpilih mengidentifikasi dua temuan utama. Pada kejadian pertama, kepatuhan kontrol pascastroke menunjukkan efek protektif yang signifikan, menurunkan risiko kekambuhan sebesar 71,1% dengan hazard ratio sebesar 0,289 dan p-value sebesar 0,035. Pada kejadian kedua, pengaruh diabetes melitus terhadap risiko kekambuhan tidak bersifat konstan, melainkan melemah secara signifikan seiring bertambahnya waktu sejak kekambuhan pertama. ======================================================================================================================================
Survival analysis is commonly used to study the time to occurrence of clinical events. In many medical studies, the same event can recur within the same individual, so the underlying risk dynamics depend on the order and spacing between events. This study applies recurrent event modeling using two time-scale frameworks, namely the Prentice-Williams-Peterson Total Time (PWP-TT) and Gap Time (PWP-GT) models. Parameter estimation was performed by maximizing the stratified partial likelihood function through the Newton-Raphson numerical iteration. The fundamental difference between the two approaches lies in the definition of the risk set within each event stratum. The results show that the Extended Cox PWP-TT model was selected as the best-fitting model based on the lowest Akaike Information Criterion (AIC) value compared to the PWP-GT model. Interpretation of the selected model was focused on the first and second event orders, while the third event order was excluded from substantive discussion due to sample size limitations (n = 11), which caused unstable parameter estimation at that stratum. The selected PWP-TT model identified two main findings. At the first event, post-stroke control adherence showed a significant protective effect, reducing the risk of recurrence by 71.1% (HR = 0.289; p = 0.035). At the second event, the effect of diabetes mellitus on recurrence risk was not constant, but weakened significantly over time since the first recurrence (time-varying component HR = 0.373; p = 0.038).

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Kejadian Berulang, PWP-TT, PWP-GT, Partial Likelihood, Stroke, Partial Likelihood, PWP-TT, PWP-GT, Recurrent Event, Stroke.
Subjects:	Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) R Medicine > R Medicine (General) > R853.S7 Survival analysis (Biometry)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Atika Nur Royyanah
Date Deposited:	05 Aug 2026 04:05
Last Modified:	05 Aug 2026 04:05
URI:	http://repository.its.ac.id/id/eprint/143958

Actions (login required)

View Item