院內泌尿道感染監視系統成效評估

System Evaluation for HAIs Surveillance System of Urinary Tract Infections

陳國斌(Guo-Bin Chen)

指導教授 ： 劉建財

臺北醫學大學/醫學科技學院/醫學資訊研究所/碩士(2012年)

https://doi.org/10.6831/TMU.2012.00182

摘要

院內感染監視在院內感染控制中，扮演著相當重要的角色，因為定期的院感監視資料分析有助於管理院內感染異常事件發生，以採取有效感染控制措施，並可作為防治措施執行成效的評估依據，提升院內感染控制品質。院內感染的定義相當複雜，需要許多檢驗數據來源以及人為的判斷，而傳統的監視方法依賴人工進行審查病人的病歷和實地病房訪查，相當耗時耗力。據研究指出感染控制護理師審查一個案約需花費 37 分 15 秒的時間。

現今醫療院所的資訊系統發展逐步成熟，目前大部分醫療服務作業已經電子化，但感染控制護理師仍需定期檢視細菌培養陽性報告日報表篩選陽性反應病人，進行院內感染之病歷檢視，雖然這些資料大部分已經電子化，但並未加以整合運用。而自動化資料收集與整合可減少感染控制人員在資料收集上的時間，自動監視系統將協助收集院內感染資料，資料的範圍以及院內感染所需要的資料作為收集標的。以院內泌尿道感染的收集資料而言，包括檢驗結果、抗生素用藥紀錄、侵入式裝置紀錄、病人住院基本資料等。然而要進行自動化必須考量到醫院的資訊系統是否可提供電子化資料，對於現存有的電子化資料進行收集。但院內感染所需的資料並非所有資料皆已電子化，例如臨床症狀，必須仰賴感控人員蒐集到病房或病歷室查閱病歷，再利用系統輸入臨床生理症狀資料。當感染控制護理師收集完生理症狀後可以利用系統來檢查現有病人資料是否符合監視標準，簡化資料收集與比對作業時間。

由於醫院需要花費大量的資源以及人力才能實施感染監視定義上的作業，雖然某醫學中心已採用醫療系統支持日常的臨床作業，但是若能使用電子自動化的系統來配合監測定義，可使感染控制人員在作業上更有效率。因此針對某醫學中心，設計一套院內感染監視系統，並上線使用。為了解某醫學中心使用之院內感染監視系統之效能與功能性，因此進行院內泌尿道感染之監視與系統執行與系統評估。

本研究目的為是否可以透過系統介入監視後，減少需要感染控制護理師人工篩選的病人數量，降低工作份量。系統是否提升感染監視偵測的準確率。使用敏感度(Sensitivity)、特異度(Specificity)、陽性預測值(Positive Predictive Value)作為系統監視院內感染作業之能力評估指標。

在實驗期間住院病人共11,321人，院內感染監視系統提示的疑似個案人數共694人，實際的院內泌尿道感染人數為93人，系統判斷的疑似個案數比實際發生個案多601個案。系統的敏感度為100 %，特異度為94.6 %，陽性預測值為13 %。

在我們的研究中，院內感染監視系統介入監視後，縮小監視的範圍，並提供完整的電子化資料，大幅降低感染控制護理師需要人工篩選的病人數量以及需查閱的病歷 資料，減少工作份量，使感染控制護理師能夠更簡單方便的進行院內感染監視作業。雖然監視的範圍縮小，但是系統偵測的敏感度還是高達100%，監測精準度相當高。因此，HAISIS可以支持感染控制專業人員更及時和有效地的日常監控作業。

雖然院內感染監視系統已經將陽性預測值從原有的4%提升至13%，但是與其他文獻比較之下，還是稍微偏低，因此未來可以針對系統的演算法做一些修改，將陽性預測值提高，使HAISIS能夠發揮更大的效能，增進院內感染監視之效率。

關鍵字

院內感染監視 ； 醫院獲得性感染 ； 監視資訊系統 ； 系統評估

並列摘要

Nosocomial infection surveillance within nosocomial infection control center, play an important role because, regular hospital infection surveillance data analysis aid in the management of abnormal nosocomial infections events. To enhance the quality of hospital infection control, effective control measures need to be taken, and evaluation of preventive measure performances.
The definition of nosocomial infection is quite complex, requiring many sources of test data and human judgment, traditional surveillance methods relying on manual review of patient records and field ward visits, which can be time-consuming and labor-intensive. Some research have pointed out that for an infection control nurse to review a case, it takes approximately 37 minutes and 15 seconds.

The development of hospital's information system have gradually matured, most of the medical service operations are electronic, but infection control professionals (ICPs) still need to regularly review the bacterial culture positive statements of the reporting date to screen positive culture patients and start a nosocomial chart review. Although this information is in electronic form, it cannot be integrated and used. Automated data collection and integration can reduce the time of the ICPs in data collection, and automatic monitoring system will help to collect nosocomial infection data, the scope of information as well as nosocomial infections in the required information to collect of the subject. Information collected on nosocomial urinary tract infection, includes positive culture results, antibiotic treatment record, the record of invasive device and inpatient’s basic information. However, in order to automate we must consider whether the hospital's information system provide electronic information for existing electronic data collection.
Required information for nosocomial infections have not all been switched to electronic format, for example such as clinical symptoms, we must rely on infection control personnel to collect the medical records from the ward or medical records room, and then use the system to enter information on clinical physiological symptoms. When the ICPs collect the physiological symptoms, they can take advantage of the system to examine existing patient data compliance monitoring standards, simplify data collection and reduce time in comparing data.
Hospitals need to spend a lot of resources and manpower to implement infection surveillance in clinical work, although some medical centers have already implemented health care system to support day-to-day clinical operations, but with the application of electronic automation system coordinating with monitoring, this will enable ICPs to be more efficient on the job. Therefore, we designed an on-line nosocomial infection surveillance systems aimed for medical center. In order to understand the performance and functionality of the nosocomial infection surveillance system in the medical center, implementation and system evaluation of nosocomial urinary tract information was executed.
The objective of this study presented here is to observe after the implementation of surveillance can it help ICPs reduce the number of manually screened patients and workload, and if the system can elevate the accuracy of monitoring and detecting infection. As an indicator for nosocomial infection systems, sensitivity, specificity and positive predictive value ability were used.

During the study period there were a total of 11,321 inpatients, nosocomial infections surveillance system detected a total of 694 suspected cases, the confirmed number of nosocomial urinary tract infection was 93. The number of suspected cases detected by the system was 601 more than actual cases. The system sensitivity was 100%, specificity was 94.6%, and the positive predictive value was 13%.

In our study, the HAISIS system involved in monitoring, narrows the scope of the monitor, and provides a complete electronic data, dramatically reducing the infection control personnel’s need to manually filter the number of patients and access medical records thus, reduction of work loading, so that infection control personnel work on simple and convenient nosocomial infection daily surveillance tasks. Although the scope of the monitor is reduced but the sensitivity of the system detection is still 100%, the monitoring of accuracy is still quite high. HAISIS can support infection control personnel’s with more timely and effective daily monitoring task.
Although HAISIS had positive predictive value from 4% to 13%, but compared with other study it is slightly lower, therefore in the future changes can be made on the algorithm of the system, to improve the positive predictive value, so that HAISIS can perform better and enhance the efficiency of the nosocomial infection surveillance.

並列關鍵字

Nosocomial infection surveillance ； Hospital-acquired infection ； surveillance information systems ； System evaluation

參考文獻

行政院衛生署疾病管制局院內感染監視通報系統.(2011). 醫療照護相關感染監視定義. 行政院衛生署疾病管制局.

行政院衛生署.(2011). 98年院內感染監視通報系統統計分析.

王復德,& 陳瑛瑛,& 陳宜君. (2010).感染管制護理師執行感管業務之時間分佈探討. 感染控制雜誌.

行政院衛生署疾病管制局.(2008). 傳染暨統計年報.

Emori, T.G., Culver, D.H., Horan, T.C. et al., (1991)National nosocomial infections surveillance system (NNIS): description of surveillance methods, Am J Infect Control.19(1), pp.19-35.