Background
All over the world, the availability of hospital beds is one of the fundamental challenges of modern healthcare. On the other hand, the COVID-19 pandemic and underlying illnesses exacerbate this problem by increasing the number of hospitalization days. Therefore, this study aims to determine the factors that affect the number of days of hospitalization of people with diabetes and with COVID-19.
Method
The Artificial neural network models used information from 183 of 200 patients to train the multilayer feed-forward back-propagation algorithms in 70% of the data. Then, model performance was assessed by the remaining 30%. In model selection, we evaluated different combinations of connection, the number of hidden layers, and the number of neurons in each hidden layer.
Results
Based on results, biorhythm cycles (physical biorhythm and emotional biorhythm) and duration of symptoms had second and third importance after demographic factors (age, hypertension, addiction, and Q2 therapy) in predicting the duration of hospitalization. And the two variables of age (22.3%) and addiction (16.6%) contribute the most to importance in the demographic section.
Conclusion
So the number of hospitalization days decreased when the disease was controlled by providing adequate training for affected individuals, conducting regular and periodic screenings for early diagnosis of mental illnesses, consuming a high-fiber, low-carb, low-fat diet, receiving proper sleep training, utilizing cognitive behavioral therapies, participating in resistance sports. Additionally, replacing long-term care centers such as nursing homes where educated nurses provide services to patients, will impose less cost on the government as compared to hospitals.