ABSTRACT
Objective To evaluate the predictive effect of Braden scale for the risk of development of pressure ulcers (PU) in the department of neurology bedridden patients and to explore subgroup preventive measures. Methods 400 cases newly hospitalized bedridden patients in the department of neurology were collected with no pressure ulcers at the first evaluation and pressure ulcer risk was continuously predicted by a Braden scale skin assessment. The high-risk, middle-risk and low-risk groups were randomized into the experimental group and the control group respectively. Routine preventive measures were taken for the control group while the air fluidized bed for the high-risk group, the sponge mattress for the middle-riskgroup, and turning the body over every 4 hours for low-risk group. Other preventive procedures were undertaken simultaneously in beth the experimental and the control groups. Results The area under the ROC curve (AUC)was 0.771 and 0.828 at the first and last time Braden scale scores respectively. Such vMues as sensitiveness, specificity, positive predictive value, negative predictive value were found in higher level,when the diagnosis value was 17. There was no significant difference of incidence rate of the subgroup pressure ulcers between the high-risk, middle-risk, low-risk groups compared to the control group. Conclusions The effect of predicting pressure ulcer risk for bedridden patients in the department of neurology with Braden scale was fairly good, while the score 17 as the diagnosis value was ideal. The air fluidized bed for the high-risk group and the sponge mattress for the middle-risk group resulted in no significant decrease of incidence rate of the pressure ulcer, while taming over the patients' body every 4 hours for low-risk groups showed acceptable and therefore saving medical resources.
ABSTRACT
GeneHub software developed by us can be applied to evaluate the significance of expression similarity of genes in each functional unit of different gene function classification systems, and thus to select experiment condition related gene functional units. With different expression similarity measurement and different experiment data, we have applied GeneHub to analyze the expression correlation of functional related genes classified respectively with gene chromosome location, celluar location and interaction, metabolic pathway and signal transduction pathway relationship of protein products of genes. Genes classified according to these multiple ways all showed significant experimental condition related co-expression in our analysis. Our results provide further evidence for the assumption that functional related genes express similarly, which is widely adopted in gene expression profile analysis.