Tesis

Model Prediksi Pneumonia 30 Hari Pasca Bedah Pada Pasien Usia Lanjut Yang Menjalani Bedah Abdomen di Rumah Sakit Cipto Mangunkusumo = Prediction Model of 30-Day Postoperative Pneumonia in Elderly Patients Undergoing Abdominal Surgery at Cipto Mangunkusumo General Hospital .

Latar Belakang. Jumlah pasien usia lanjut yang menjalani pembedahan semakin meningkat. Pasien yang menjalani bedah abdomen dan mengalami pneumonia pasca bedah akan mengalami peningkatan risiko terjadinya mortalitas dibandingkan mereka yang tidak mengalami pneumonia pasca bedah. Penelitian mengenai faktor risiko terjadinya pneumonia pasca bedah pada pasien usia lanjut yang menjalani bedah abdomen masih sedikit. Belum ada model prediksi pneumonia pasca bedah pada pasien usia lanjut yang menjalani bedah abdomen. Tujuan. Mendapatkan model prediksi pneumonia 30 hari pasca bedah pada pasien usia lanjut yang menjalani bedah abdomen. Metode. Penelitian dengan desain kohort retrospektif dilakukan terhadap pasien usia lanjut yang menjalani bedah abdomen di Rumah Sakit Cipto Mangunkusumo pada periode waktu Januari 2013 sampai Desember 2017. Prediktor independen pneumonia 30 hari pasca bedah diidentifikasi dengan analisis regresi logistik. Performa kalibrasi dinilai dengan uji Hosmer-Lemeshow. Performa diskriminasi dinilai dengan area under the curve (AUC). Hasil. Empat prediktor independen pneumonia 30 hari pasca bedah diidentifikasi, yakni status fungsional (ADL Barthel 0-11 , Odds Ratio (OR) 6,908 [ Interval Kepercayaan /IK 95 % 2,933-16,273]; ADL Barthel 12-19, OR 3,191 [ IK 95 % 1,53-6,657]), bedah abdomen atas ( OR 4,869 [IK 95% 1,805-13,132]), PPOK (OR 6,88 [ IK 95 % 2,00123,709]) dan albumin < 3 g/dl (OR 2,54 [IK 95 % 1,404-4,596]). Dari uji Hosmer – Lameshow didapatkan p 0,452 dan AUC 0,811 [ IK 95 % 0,75-0,86]). Kesimpulan. Model prediksi terjadinya pneumonia 30 hari pasca bedah pada pasien usia lanjut yang menjalani bedah abdomen dapat disusun berdasarkan empat prediktor yaitu status fungsional, bedah abdomen atas, PPOK dan kadar albumin < 3 g/dl. Model prediksi tersebut memiliki diskriminasi dan kalibrasi yang baik.
Kata Kunci. usia lanjut, bedah adomen, pneumonia pasca bedah, model prediksi


Background. The number of elderly patients who undergo surgery are increasing. Those who had abdominal surgery and developed postoperative pneumonia have higher risk of mortality compared to those who did not. Study about the risk factors of postoperative pneumonia in elderly patients undergoing abdominal surgery are still limited. The prediction model for postoperative pneumonia in elderly patients who undergo abdominal surgery is not available yet. Objective. To develop a prediction model of 30-day postoperative pneumonia in elderly patients undergoing abdominal surgery. Methods. A retrospective cohort study was conducted in elderly patients who went abdominal surgery in Cipto Mangunkusumo General Hospital from January 2013 to December 2017. Independent predictors for 30-day postoperative pneumonia were identified by logistic regression analysis. Calibration performance of the model was tested by Hosmer-Lameshow test. Discrimination performance of the model was tested by calculating area under curve (AUC). Results. Four independent predictors for 30-day postoperative pneumonia were identified, including : functional status (ADL Barthel 0-11, Odds Ratio (OR) 6.908 [ 95 % Confidence Interval /CI 2.933-16.273]; ADL Barthel 12-19, OR 3.191 [ 95 % CI 1.53-6.657]), upper abdominal surgery ( OR 4.869 [95% CI 1.805-13.132]), COPD (OR 6.888 [ 95 % CI 2.001-23.709]) and albumin < 3 g/dl (OR 2.54 [95 % CI 1.404-4.596]). The Hosmer-Lemeshow test revealed p-value 0.452 and the AUC value is 0.811 [95% CI 0.87-0.97]). Conclusion. A prediction model of 30-day postoperative pneumonia for elderly patients undergoing abdominal surgery can be developed based on four predictors, i.e. functional status, upper abdominal surgery, COPD and albumin level < 3 g/dl. The prediction model has good discrimination and calibration.
Keywords. elderly, abdominal surgery, postoperative pneumonia, prediction model.