Rachmawati, Juwitasari Nur (2019) Propensity Score Matching (PSM) Menggunakan Metode Classification Tree pada Kasus HIV/AIDS di Puskesmas Grati Kabupaten Pasuruan. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
06211540000046_Undergraduate_Theses.pdf - Published Version Download (2MB) | Preview |
Abstract
Kejadian Infeksi Oportunistik (IO) menjadi penyebab utama kematian pada lebih dari 90% penderita HIV/AIDS. Infeksi Oportunistik adalah infeksi yang timbul akibat penurunan kekebalan tubuh yang disebabkan oleh HIV/AIDS dan mendapatkan perhatian yang luas baik di Indonesia maupun dunia. Penelitian ini menggunakan data pasien HIV/AIDS yang dirawat di Puskesmas Grati Kabupaten Pasuruan yang terdiri dari 7 variabel prediktor (X) dan variabel kejadian IO (Y). Tujuan penelitian ini adalah mengatasi estimasi yang bias pada efek perlakuan yang disebabkan oleh penelitian observasional dan menghitung persentase bias tereduksi. Metode yang digunakan dalam analisis ini adalah Propensity Score Matching (PSM) dengan metode estimasi nilai propensity menggunakan Classification Tree. Ditemukan bahwa tipe pemberian terapi merupakan variabel confounding dalam penelitian ini. Hasil analisis PSM menunjukkan bahwa ada sebanyak 36 data pasien yang diberikan terapi ARV dan pendampingan dipasangkan dengan pasien yang diberikan terapi ARV. Lalu dilakukan evaluasi PSM yakni menggunakan Average Treatment of Treated (ATT), ditunjukkan bahwa variabel confounding setelah dilakukan analisis PSM berpengaruh signifikan terhadap respon. Metode PSM dengan menggunakan covariate yang sudah balance pada penelitian ini mampu mereduksi bias yang ditimbulkan dari kasus studi observasional sebesar 38,61%.
==============================================================================================================================
Oportunistic Infection (OI) still being the main cause of death for more than 90% of HIV/AIDS patients. Opportunistic Infection is an infection caused by decreased immunity of HIV/AIDS and now become the main concern for the health sector both in Indonesia and abroad. This research uses HIV/AIDS patients data whom treated at Puskesmas Grati, Pasuruan District that consists of 7 covariates (X) and 1 response variable which is the OI event (Y). The main purpose of this thesis is to resolve bias estimation of confounding variable that caused by non-random observation, and another goal is to calculate the Percent Bias Reduction (PBR). The method used in this research is Propensity Score Matching (PSM) using Classification Tree Approach to estimate the propensity score. It is found from the research that the type of theraphy given to the patients become the confounding variable. The result shows that there are 36 data of patients that were given ARV and consultation theraphy match with the data of patients that were given ARV theraphy. Then PSM is being evaluated from the value of Average Treatment of Treated (ATT), and it gave result that the confounding variable after being processed by PSM is significantly affecting response variable. The PSM method using the covariate which were balance in this research, is proven to be successfully reduce the bias that caused by non-random observation for about 38,61%.
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSSt 519.536 Rac p-1 2019 |
| Uncontrolled Keywords: | Classification Tree, HIV/AIDS, IO, PSM |
| Subjects: | Q Science Q Science > QA Mathematics Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA76.6 Computer programming. Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics) |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Juwitasari Nur Rachmawati |
| Date Deposited: | 26 Dec 2022 02:57 |
| Last Modified: | 26 Dec 2022 02:57 |
| URI: | http://repository.its.ac.id/id/eprint/63392 |
Actions (login required)
 |
View Item |