Heart rate characteristics monitoring in newborn infants

描述(申请人提供)：败血症是中国的一个主要健康问题 新生儿重症监护病房(NICU)的高危新生儿，在那里 发生在25%的极低出生体重儿中，并导致超过 死亡率增加一倍，住院时间增加50%。我们的长期合作 目的是检验这样一种假设，即检测异常心率 具有持续非侵入性监测的特征(HRC)将改善护理 在这些患者中，通过早期诊断为脓毒症和其他亚急性， 潜在的灾难性疾病。我们建议朝着这一目标前进 通过完成两个研究目标。 在目标1中，我们将前瞻性地研究大学NICU中未被选中的婴儿。 测试异常的HRC将与即将到来的 客观疾病标准定义的败血症和脓毒症样疾病；以及 人权委员会增加了关于以下风险的重要独立诊断信息 脓毒症和脓毒症样疾病对出生体重和临床变量的影响 胎龄。临床研究设计是为了派生出 跟踪一组患者的多变量预测统计模型 通过使用来自第二组患者的临床数据的验证阶段。 在目标2中，我们将制定针对以下任务的人权委员会新措施 早期发现新生儿败血症。我们将调查最佳选择 样本熵计算参数、心率时间的平稳性 使用基于经验累积的传统测量和新测量的系列 分布函数，并使用一种新方法进行频域分析。

DESCRIPTION (provided by applicant): Sepsis is a major health problem in high-risk newborn infants in the neonatal intensive care unit (NICU), where it occurs in 25 percent of very low birth weight infants and leads to a more than doubling of mortality and a 50 percent increase in hospital stay. Our long-term objective is to test the hypothesis that detection of abnormal heart rate characteristics (HRC) with continuous non-invasive monitoring will improve care of these patients by earlier diagnosis of sepsis and other sub-acute, potentially catastrophic illnesses. We propose to advance toward this objective by completing two research aims. In Aim 1, we will prospectively study unselected infants in a university NICU to test the hypotheses that abnormal HRC will be associated with upcoming sepsis and sepsis-like illness as defined by objective illness criteria, and that HRC adds significant independent diagnostic information about the risk of sepsis and sepsis-like illness to clinical variables of birth weight and gestational age. The clinical research design is for derivation of multivariable predictive statistical models from one set of patients followed by a validation phase using clinical data from a second set of patients. In Aim 2, we will develop new measures of HRC that are specific to the task of early detection of neonatal sepsis. We will investigate optimum choice of parameters for sample entropy calculation, stationarity of heart rate time series using conventional and novel measures based on the empirical cumulative distribution function, and frequency domain analysis using a novel method.