PENERAPAN DATA MINING UNTUK IDENTIFIKASI POLA PERESEPAN OBAT PADA PASIEN DIABETES MELLITUS DAN PENYAKIT KOMPLIKASINYA DI RSUD DUMAI PROVINSI RIAU

TRI WIDAYANTI, Prof. dr. Hari Kusnanto, DrPH ; Dr. Satibi, M.Si, Apt

2015 | Tesis | S2 Ilmu Kesehatan Masyarakat

LatarBelakang: Diabetes mellitus merupakanpenyakit yang banyakdideritaolehpendudukduniasertadapatmenyebabkanmunculnyaberbagaipenyakitbaruataukomplikasisehinggamembutuhkanbanyakperhatian para klinisi. Hipertensimerupakanpenyakit yang paling banyakditemukansebagaipenyakitpenyerta diabetes mellitus.Penerapandata mininguntukidentifikasipolaperesepanobatpadapasien diabetes mellitus danpenyakitkomplikasinya, yaituhipertensiesensial (I10), diharapkandapatmemberikangambarangunamembantupengambilankeputusandalammemperbaikipelayananpasien diabetes mellitus sehinggadapatmengurangikematianmaupunresiko yang disebabkanolehpenyakit diabetes mellitus tersebut. Tujuan:Mengembangkankandata mininguntukidentifikasipolaperesepanobatpadapasien diabetes mellitus danpenyakitkomplikasinya, yaituhipertensiesensial (I10), di RSUD DumaiProvinsi Riau. Metode:Penelitianinimenggunakanmetodeobservasionaldeskriptifdenganrancanganpenelitiancross sectional study. Analisis yang digunakanadalahdeskriptifdenganmenggunakantehnikclustering data mining danalgoritmaK-Means. Hasil:.Persentasedarijumlahkunjungan yang diberikanobatantidiabetikadalahsebesar 91,51% danantihipertensiadalah 90,72%. Golonganobatantidiabetik yang digunakanadalah analog insulin (18,30%), biguanida (71,62%), sulfonilurea (53,85%), tiazolidinedion (31,03%), dan alfa glukosidase (0,53%), sedangkanogolanganobatantihipertensi yang digunakanadalah inhibitor ACE (34,75%), ARB (21,48%), deuretik (30,24%), beta blockers (3,45%, dan CCB (5597%). Terdapat 3 polaclustering (pengelompokan) padaperesepanobatantidiabetikdan 4 polaclustering padaperesepanobatantihipertensi, yang menggambarkanrekomendasidalamperesepanobatpadapasien diabetes mellitus dengandisertaihipertensiesensialgunamendapatkanhasilterapi yang lebihbaik. Kesimpulan:Polaclustering dariperesepanobatantidiabetikdanobatantihipertensiinirelevanuntukpengambilankeputusankliniskarenamemberikangambaranperesepan yang direkomendasikandantelahdiakuiolehbeberapapenelitian. Kata kunci:clustering, data mining, diabetes mellitus, hipertensiesensial, K-Means, peresepanobat.

Background: Diabetes mellitus have been a disease that suffered by many population in the world and able to cause the emerging of various complications. The complications that most often be found in diabetes mellitus was hypertension.The implementation of data miningto identifying the prescription pattern of medicine for the diabetes mellitus patients and their complication diseases that was essential hypertension, was expected able to provide a depiction in order to aid a decision making to improve a service for diabetes mellitus patients thus able to reduce both mortality and a risks caused by diabetes mellitus. Objectives: To develop data miningto identifying a prescription pattern of medicine for diabetes mellitus patients and their complication diseases or illness that is essential hypertension in Dumai Regional General Hospital, Riau Province. Methods: This study used descriptive observational method by cross sectional design. Results: The group of antidiabetic medicine that were used is insulin analog (18.30%), biguanida (71.62%), sulfonylurea (53.85%), tiazolidinedion (31.03%), and alfa glucosidase (0.53%), whereas a group of antihypertension medicine that were used was inhibitor ACE (34.75%), ARB (21.48%), diuretic (30.24%), beta blockers (3.45%) and CCB (55.97%). There were 3 pattern of clustering (grouping) in the prescription of antidiabetic medicine and 4 clustering pattern in the prescription of antihypertension medicine, that depicted a recommendation in the medicine prescription for diabetes mellitus patients who with essential hypertension in order to get better therapy outcome. Conclusion:This clustering pattern of the prescription of antidiabetic medicine and antihypertension medicine is relevant for clinical decision making because provide a depiction of prescriptions that were recommended and already recognized by several studies. Keywords: clustering,data mining, diabetes mellitus, essential hypertension, K-Means, medicine prescription

Kata Kunci : clustering,data mining, diabetes mellitus, essential hypertension, K-Means, medicine prescription