Pak Changsik, In Jeon Ji, Kim Hyeonwoo, Kim Jungyoon, Park Suyeon, Ahn Ki-Hwan, Son Yeon-Joo, Yoo Sooyoung, Baek Rong-Min, Jeong Jae Hoon, Heo Chan Yeong
Department of Plastic and Reconstructive Surgery, Seoul National University Bundang Hospital, Gyeonggi, Republic of Korea.
Department of Technology Development, KT R&D Center, Seoul, Republic of Korea.
Wound Repair Regen. 2018 Dec;26 Suppl 1:S19-S26. doi: 10.1111/wrr.2.
We investigated the accuracy of pressure injury evaluation using tele-devices and examined the concordance between automatically generated recommendations and primary manual recommendations. Caregivers took photos and videos of pressure injuries using smartphones with built-in cameras and uploaded the media to the application. The wound team evaluated the wound using a specially modified version of the Pressure Sore Status Tool. This was compared with the Pressure Sore Status Tool score assessed during the actual examination of the patient. We developed an automatic algorithm for dressing based on the Pressure Sore Status Tool score, checking for consistency between this and the primary manual recommendation. A total of 60 patients diagnosed with pressure injuries were included. The κ coefficients indicated substantial agreement for wound size and total score, and excellent for all other items. We found that the overall concordance rates were statistically significant for all items (p < 0.001). For the primary dressing, the κ coefficient for the concordance rate of automatic algorithm and manual recommendation was 0.771, while that of teleconsultation system and manual recommendation was 0.971. For the secondary dressing, the figures were 0.798 and 0.989, respectively. All values were statistically significant (p < 0.001). We presented strong evidence documenting the utilization of a smartphone, patient-driven system, and demonstrated that the measurements obtained were comparable to the ones obtained by a trained, on-site, wound team. Furthermore, we confirmed agreement between automatically generated recommendations and primary manual recommendations.
我们研究了使用远程设备进行压力性损伤评估的准确性，并检查了自动生成的建议与初步人工建议之间的一致性。护理人员使用内置摄像头的智能手机拍摄压力性损伤的照片和视频，并将这些媒体文件上传到应用程序中。伤口护理团队使用经过特殊修改的压力性溃疡状态工具对伤口进行评估。将其与在对患者进行实际检查期间评估的压力性溃疡状态工具评分进行比较。我们基于压力性溃疡状态工具评分开发了一种自动换药算法，并检查其与初步人工建议之间的一致性。总共纳入了60例被诊断为压力性损伤的患者。κ系数表明在伤口大小和总分方面有实质性一致性，在所有其他项目上一致性良好。我们发现所有项目的总体一致性率在统计学上具有显著性（p < 0.001）。对于初次换药，自动算法与人工建议的一致性率的κ系数为0.771，而远程会诊系统与人工建议的κ系数为0.971。对于二次换药，相应的数字分别为0.798和0.989。所有值在统计学上均具有显著性（p < 0.001）。我们提供了有力证据证明了使用患者驱动的智能手机系统的情况，并表明所获得的测量结果与由训练有素的现场伤口护理团队获得的结果相当。此外，我们证实了自动生成的建议与初步人工建议之间的一致性。
