PURPOSE: The objective of this study was to develop and evaluate a computerized rule-based decision support algorithm for nursing triage of potential acute bronchopulmonary events in lung transplant recipients on the basis of home monitoring of spirometry and symptoms. METHODS: The algorithm automatically separates recipients into 2 groups: those who are stable or improving and those who should be "watched" further because of their potential for developing bronchopulmonary events according to their weekly home-monitoring measurements. A total of 155 recipients (82 females and 73 males) contributed 1944 weekly records for the training (420), testing (786), and prospective evaluation (738) data sets. Weekly records contained daily values of forced expiratory volume at 1 second and respiratory symptoms, which were the inputs for the triage rules. RESULTS: Sensitivity and specificity were greater than 90% for the prospective evaluation comparing the computer decision support system with the manual nurse review of the same home-monitoring reports. CONCLUSIONS: Algorithm performance in identifying lung recipients who should be placed on a "watch" list for the potential to develop acute bronchopulmonary events is comparable to the standard human clinical review of the same weekly home-monitoring data.
|Original language||English (US)|
|Number of pages||8|
|Journal||Heart and Lung: Journal of Acute and Critical Care|
|State||Published - May 2005|
Bibliographical noteFunding Information:
Supported in part by National Institutes of Health Grant RO1 NR02128.