Survey of Electronic Health Records Data for Developing a Predictive Model of Pressure Ulcers in Critical Care Patients
| Main Author: | |
|---|---|
| Language: | English |
| Published: |
The Ohio State University / OhioLINK
2012
|
| Subjects: | |
| Online Access: | http://rave.ohiolink.edu/etdc/view?acc_num=osu1338371919 |
| id |
ndltd-OhioLink-oai-etd.ohiolink.edu-osu1338371919 |
|---|---|
| record_format |
oai_dc |
| spelling |
ndltd-OhioLink-oai-etd.ohiolink.edu-osu13383719192021-08-03T06:05:34Z Survey of Electronic Health Records Data for Developing a Predictive Model of Pressure Ulcers in Critical Care Patients Panchagavi, Renuka Computer Engineering Electronic Health Records Pressure Ulcers Classification Pressure Ulcers are a hospital-acquired condition hindering functional recovery in patients and resulting in additional cost[5]. Although pressure ulcers are preventable, the prevalence of pressure ulcers in health care facilities is still not satisfactory[5]. Critically ill patients in the intensive care unit (ICU) are at particularly high risk of pressure ulcers. This has sparked an urgency and awareness related to constructing an ideal preventive system to improve the quality of patient care. Accurate identification of the risk factors for pressure ulcers can lead to reductions in both the occurrence of pressure ulcers and health care costs and can promote positive health outcomes in critical care patients. The objective of this work is to mine the large and heterogeneous Electronic Health Records (EHR) data of adult ICU patients by applying data mining techniques, and thereby identify important factors accountable for pressure ulcers in ICU patients, that will be useful for prediction modeling of Pressure Ulcers among patients in intensive care setting, and use AI methods such as, Decision Trees (DT) to construct a feasible prediction model for pressure ulcers in ICU patients, using WEKA, an open source software package[10]. The parameters used for evaluating classification accuracy of the Decision Tree Classifier include TP Rate, FP Rate, Precision, Recall, F-Measure and AUC (Area under the ROC curve). This study also focuses on measuring statistical significance of association between medications and risk of pressure ulcer development in ICU patients, using the Fisher’s Exact Test and also uses Bi-clustering method to identify clusters enriched with patients who developed pressure ulcers and thereby find the enriched categories of drugs . The results show that a feasible prediction model can be constructed , based on demographic attributes like patient’s age, gender, and length of stay as features, in order to predict pressure ulcer development in ICU patients. Results indicate that, length of stay proved to be better predictor of pressure ulcer development than the other 2 features, gender and age. According to the Fisher’s Exact test results, the drugs that were found to be significant were Electrolytes, Sedation drugs, Diuretics, Cardiac drugs and few antibiotics and anti-fungal drugs. Based on the summarized results of bi-clustering, eight clusters were identified as enriched with patients who developed pressure ulcers and the enriched categories of drugs were found to be – Sedatives, IV Fluids and Diuretics. Thus, we conclude that an association study of clinical variables from EHR data of adult ICU patients will provide a useful insight towards assessing a patient’s risk for developing a pressure ulcer and thus, can help reduce the associated costs for treating pressure ulcers. 2012-06-26 English text The Ohio State University / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=osu1338371919 http://rave.ohiolink.edu/etdc/view?acc_num=osu1338371919 unrestricted This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws. |
| collection |
NDLTD |
| language |
English |
| sources |
NDLTD |
| topic |
Computer Engineering Electronic Health Records Pressure Ulcers Classification |
| spellingShingle |
Computer Engineering Electronic Health Records Pressure Ulcers Classification Panchagavi, Renuka Survey of Electronic Health Records Data for Developing a Predictive Model of Pressure Ulcers in Critical Care Patients |
| author |
Panchagavi, Renuka |
| author_facet |
Panchagavi, Renuka |
| author_sort |
Panchagavi, Renuka |
| title |
Survey of Electronic Health Records Data for Developing a Predictive Model of Pressure Ulcers in Critical Care Patients |
| title_short |
Survey of Electronic Health Records Data for Developing a Predictive Model of Pressure Ulcers in Critical Care Patients |
| title_full |
Survey of Electronic Health Records Data for Developing a Predictive Model of Pressure Ulcers in Critical Care Patients |
| title_fullStr |
Survey of Electronic Health Records Data for Developing a Predictive Model of Pressure Ulcers in Critical Care Patients |
| title_full_unstemmed |
Survey of Electronic Health Records Data for Developing a Predictive Model of Pressure Ulcers in Critical Care Patients |
| title_sort |
survey of electronic health records data for developing a predictive model of pressure ulcers in critical care patients |
| publisher |
The Ohio State University / OhioLINK |
| publishDate |
2012 |
| url |
http://rave.ohiolink.edu/etdc/view?acc_num=osu1338371919 |
| work_keys_str_mv |
AT panchagavirenuka surveyofelectronichealthrecordsdatafordevelopingapredictivemodelofpressureulcersincriticalcarepatients |
| _version_ |
1719430720794918912 |