Tesis

Model Prediksi Mortalitas 72 Jam Pasien Ketoasidosis Diabetikum di Instalasi Gawat Darurat RSUPN Cipto Mangunkusumo = Seventy Two Hours Mortality Prediction Model in Diabetic Ketoacidosis Patients at Emergency Unit of Cipto Mangunkusumo General Hospital.

Latar Belakang. Mortalitas KAD sebagai komplikasi akut DM di negara berkembang seperti Indonesia masih tinggi. Karena itu, diperlukan model prediksi untuk menapis pasien-pasien KAD yang memiliki risiko mortalitas tinggi. Tujuan. Mendapatkan model prediksi mortalitas 72 jam pasien KAD di IGD RSUPN Cipto Mangunkusumo. Metode. Penelitian dengan desain kohort retrospektif menggunakan rekam medik pasien KAD di IGD RSUPN Cipto Mangunkusumo periode Januari 2011 – Juni 2017 dengan metode sampling konsekutif. Hubungan mortalitas 72 jam dengan prediktor yaitu, usia, tingkat kesadaran, jenis DM, riwayat KAD, jumlah komorbid dan parameter laboratorium (kadar bikarbonat, kalium, anion gap, βhidroksibutirat, laktat dan fungsi ginjal) akan dinilai dan dilanjutkan dengan pembuatan model prediksi. Seluruh analisis dilakukan menggunakan program SPSS Statistics 20.0. Hasil. Sebanyak 86 subjek (28,57%) dari 301 subjek yang dianalisis meninggal dalam waktu 72 jam. Prediktor yang berhubungan bermakna dengan mortalitas pada analisis multivariat (p 4 mmol/L (HR 5,585; IK 95% 2,966 – 10,519). Keempat prediktor dilanjutkan ke dalam sistem skor dan didapatkan model prediksi mortalitas 72 jam pasien KAD RSCM yang memiliki performa baik dengan probabilitas mortalitas sebesar 15,41 % untuk skor 0 – 2, 78,01% untuk total skor 3 – 4 dan 98,22% untuk total skor 5 – 6. Simpulan. Mortalitas 72 jam pasien KAD di RSUPN Cipto Mangunkusumo adalah 28,57%. Model prediksi mortalitas memiliki performa yang baik dan terdiri dari komorbid, riwayat KAD, tingkat kesadaran, dan kadar laktat.
Kata Kunci. Model Prediksi, Mortalitas, Ketoasidosis Diabetikum


Background. Mortality rate of DKA as acute complication of DM in Indonesia is still high. Therefore, a mortality prediction model is needed to screen high risk mortality DKA patients. Aim. To identify prediction model of 72 hours mortality in diabetic ketoacidosis patients at emergency unit Cipto Mangunkusumo General Hospital. Methods. This was a retrospective cohort study with consecutive sampling method. Subjects were adult DKA patients in emergency unit Cipto Mangunkusumo General Hospital from January 2011 to June 2017. Data were obtained from medical records. Association of predictors (age, type of DM, history of DKA, comorbidities, level of consciousness, bicarbonate, potassium, anion gap, lactate, β-hydroxybutirate and renal function) and 72 hours mortality was analyzed and submitted to prediction model. All analysis was done using SPSS Statistics 20.0. Results. A total of 86 subjects out of 301 subjects did not survive in 72 hours since hospital admission. Comorbidities (HR 2,407; 95% IC 1,181 – 4.907), level of consciousness (HR 10,345; 95% IC 4,860 – 22,019), history of DKA (HR 2,126; 95% IC 1,308 – 3,457) and lactate level (HR 5,585; 95% IC 2,966 – 10,519) were significant predictors and submitted to scoring system. A prediction model was derived with a good performance. Subjects with 0 – 2 points were at 15,41 % risk of mortality, 3 – 4 points were 78,01% and 5 – 6 points were 98,22% risk of mortality. Conclusion. Seventy two hours mortality rate in Cipto Mangunkusumo General Hospital was 28,57%. The mortality prediction model had a good performance, consisted of comorbidities, history of DKA, level of consciousness dan lactate level.
Keywords. Prediction Model, Mortality, Diabetic Ketoacidosis