Kalp Yetersizliği Olan Hastaların Hastaneye Yeniden Yatışı İle İlgili Faktörlerin Veri Madenciliği Teknikleri İle İncelenmesi
Özet
The high levels of rehospitalization is a sign of deficient providence of the necessary health
services at the hospitals. Unplanned rehospitalization cause unnecessary expenses in the health
sector and also impact the quality of the patient negatively. Due to these reasons, the health
managers have been trying to develop policies to reduce those unplanned rehospitalizations.
With this study, the prediction of the factors that cause the rehospitalization of the patients who
had to be rehospitalized at the Health Sciences University Ümraniye Training and Research
Hospital within 30 days of being discharged by using data mining application. The researcher
has worked with the data of a total of 400 people composed of 200 patients who had been
hospitalized at the Ümraniye Training and Research Hospital between the dates of 29.08.2008
and 10.04.2014 and had to be rehospitalized within 30 days after their discharge from the
hospital, and 200 patients who had not been rehospitalized within 30 days after their discharge
from the hospital. The patient background data used in the study are the following: Age, gender,
their level of calcium, urea, creatinine, potassium and sodium, comorbidity, their number of
outpatient clinic visits in the last year, and number of emergency service visits.
The application section of this study has been made with the Weka Program which is one of the
package programs used in relation to machine learning. The model has been formed by the J.48
Algorithm of the decision tree which is one of the classification methods of data mining. According
to the prediction of the created model, the number of correct predictions out of 400 is 299
(74.75%). The number of correct predictions with the rehospitalized 200 patients is 147, and with
the non-rehospitalized patients is 152. The values of validity and reliability of the model is found
to be 74.8%.