| Summary: | 碩士 === 龍華科技大學 === 資訊管理系碩士班 === 101 === People are living longer because of advances in medical and health technology, and elderly population gradually increased. With the effects of low birth rate, the number of elderly living in long-term care institutions was gradually increased. However, the number of nurses in institutions is fixed, so how to use information technology to assist them to monitor the vital signs of the elderly is a worth discussing topic. Furthermore, it is lack to set the earninig warning range of the elderly’s vital signs that meet their personal situation, and there are few papers using the control charts in the research of the earninig warning range of the individual patients’ vital signs. Therefore, this study used the univariate and multivariate control chart techniques for long-term care patients to look for a variety of the warning range of their vital signs, and proposed the earning warning methods for the individual patient’s vital signs.
This study proposed the whole and individual health interval for the four vital signs such as temperature, pulse, respiration and blood pressure in the univariate analysis. The whole interval is divided into normal human interval and the subjective long-term care elderly’s interval. And the individual interval is divided into personal interval and statistical process control (SPC) interval. The normal human intervals are referenced from the literatures. The subjective long-term care elderly’s intervals are set by adjusting the normal human intervals in accordance with the average of the experimental samples. The personal interval is set by looking for a suit one from the individual's averages plus or minus a few standard deviations. The last SPC interval is set by using SPC software to calculate the individual values and moving range to find the individual healthy interval. According to the four healthy intervals described above, this study sets three alarm interval colors of red (R), yellow (Y) and green (G) in accordance with the severity.
In the multivariate analysis, this study used the Hotelling T2 control chart method to consider three vital signs such as body temperature, pulse and blood pressure at the same time. This study sets the multivariate early warning by using the upper limit identified from the Hotelling T2 control chart and the orange (O) warning light is truned on when the observed value exceeds the upper limit. In practical operation, this study suggests the automated sliding warning interval that uses the concept of the time window to calculate the next warning intervals. When the system is moinitoring the patients’ vital signs, it will fine-tune the warning intervals automaticly and dynamicly. So adjusting the personal health interval automaticly and dynamically can be better to meet the demand of setting the intervals according to the patient's condition, and it can also reduce the nurses’ workload. In the future, we can use the artificial neural networks or other data mining techniques to analyze the impact of age, gender and disease for vital signs to find out some groups, and then use our early warning method to find more suitable intervals of these different groups.
|
|---|
