Automated Pressure Injury Risk Assessment System Incorporated Into an Electronic Health Record System. Issue 6 (November 2017)

• Record Type: Journal Article
• Authors: Jin, Yinji; Jin, Taixian; Lee, Sun-Mi
• Is Part Of: Nursing research. Volume 66:Issue 6(2017)
• Publication Date: 2017-11
• Subjects: electronic health record -- hospital -- pressure injury -- prevention -- risk assessment -- validity; Nursing -- Research -- Periodicals; Nursing -- Periodicals; Nursing -- Periodicals; Soins infirmiers -- Recherche -- Périodiques; Soins infirmiers -- Périodiques; Verpleegkunde; Nursing; Nursing -- Research; Periodicals; 610.73
• DOI: 10.1097/NNR.0000000000000245
• Languages: English
• ISSNs: 0029-6562

Abstract

Abstract : Background: Pressure injury risk assessment is the first step toward preventing pressure injuries, but traditional assessment tools are time-consuming, resulting in work overload and fatigue for nurses. Objectives: The objectives of the study were to build an automated pressure injury risk assessment system (Auto-PIRAS) that can assess pressure injury risk using data, without requiring nurses to collect or input additional data, and to evaluate the validity of this assessment tool. Methods: A retrospective case–control study and a system development study were conducted in a 1, 355-bed university hospital in Seoul, South Korea. A total of 1, 305 pressure injury patients and 5, 220 nonpressure injury patients participated for the development of a risk scoring algorithm: 687 and 2, 748 for the validation of the algorithm and 237 and 994 for validation after clinical implementation, respectively. A total of 4, 211 pressure injury-related clinical variables were extracted from the electronic health record (EHR) systems to develop a risk scoring algorithm, which was validated and incorporated into the EHR. That program was further evaluated for predictive and concurrent validity. Results: Auto-PIRAS, incorporated into the EHR system, assigned a risk assessment score of high, moderate, or low and displayed this on the Kardex nursing record screen. Risk scores were updated nightly according to 10 predetermined risk factors. The predictive validity measures of the algorithm validation stage were as follows: sensitivity = .87, specificity = .90, positive predictive value = .68, negative predictive value = .97, Youden index = .77, and the area under the receiver operating characteristic curve = .95. The predictive validity measures of the Braden Scale were as follows: sensitivity = .77, specificity = .93, positive predictive value = .72, negative predictive value = .95, Youden index = .70, and the area under the receiver operating characteristic curve = .85. The kappa of the Auto-PIRAS and Braden Scale risk classification result was .73. Discussion: The predictive performance of the Auto-PIRAS was similar to Braden Scale assessments conducted by nurses. Auto-PIRAS is expected to be used as a system that assesses pressure injury risk automatically without additional data collection by nurses.