Model Matematika Sebagai Kecerdasan Buatan dalam Memprediksi Lama Rawatan Pasien Diabetes Melitus dengan Hipoglikemia

Zufry, Hendra
Sucipto, Krishna Wardhana
ekadamayanti, agustia sukri

Journal of Medical Science; Jurnal Ilmu Medis Rumah Sakit Umum Daerah dr. Zainoel Abidin, Banda Aceh
Vol 2 No 2 (2021): Journal of Medical Science

Background: Diabetes Mellitus (DM), a burden disease that has contributed to the health burden globally. The most common acute complication is hypoglycemia. Hypoglycemia in hospitalized diabetic patients is associated with increased length of stay (LOS), mortality, and costs. Patients with co-morbidities are often not in accordance with the clinical pathway, so a mathematical model is needed which estimates length of stay and the required resources can be estimated. Aim: To formulate a mathematical model predicting the LOS of hypoglycemic patients treated with various conditions Methode: This study is a retrospective cohort study using secondary data of DM patients who experienced hypoglycemia from January 1, 2011 to December 2020 which obtained from the medical records of Zainoel Abidin Hospital Banda Aceh, Indonesia. Data collected from May to November 2021. The data were analyzed using multivariate analysis with logistic regression test. using R software to obtain a mathematical modeling of length of stay. The variables that affect the length of stay using logistic regression test with p value <0.05. Result: There were 573 hypoglycemic patients, most of the patients who experienced hypoglycemia were male (51.3 %), the average age was 59.6 years. Most of the patients who experienced hypoglycemia were type 2 DM patients (96.4%), Infection was the main comorbid (46.6 %), Mild hypoglycemia was the highest (47.2%) and the average length of stay was 8.06 days. Based on mathematical modeling, it was found that for every additional unit of age, the length of treatment increased by 6.99 7 days. Then a logistic regression test was carried out to see the influential variables. Conclusion: Based on the mathematical modeling obtained, it is known that for each additional unit of age, the length of stay will increase by about 6.9≈7 days. Furthermore, the formulation of this mathematical model is expected to be able to predict the length of stay for various comorbidities, predict the number of consumable materials and as a basis for making artificial intelligence applications that is beneficial to efficiency.