A Multimodal Approach for Monitoring Prolonged Acute Pain in Neonates

监测新生儿长期急性疼痛的多模式方法

• 批准号: 10218273
• 负责人: Yu Sun
• 金额: $ 21.88万
• 依托单位: UNIVERSITY OF SOUTH FLORIDA
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-07-16 至 2023-06-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10218273
• 关键词: Acute Pain; Affect; Agitation; Analgesics; Appearance; Behavioral; Belief; Brain; Caregivers; Caring; Chronic Disease; Classification; Clinical Research; Communities; Computational algorithm; Computer Assisted; Computer Vision Systems; Crying; Data Collection; Descriptor; Development; Developmental Disabilities; Diagnosis; Discipline of Nursing; Documentation; Emotions; Enrollment; Ensure; Evaluation; Expert Systems; Exposure to; Face; Facial Expression; Faculty; Failure; Family; Frequencies; General Hospitals; Goals; Health; Hospitalization; Human; Incidence; Individual; Infant; Infant Care; Lead; Life; Machine Learning; Medicine; Methods; Mission; Modality; Modeling; Monitor; Motion; Neonatal; Neonatal Intensive Care Units; Neonatal Nursing; Nervous System Trauma; Notification; Nurses; Nursing Research; Opiate Addiction; Optics; Outcome; Output; Pain; Pain Assessment Tool; Pain Measurement; Pain management; Paper; Pattern; Performance; Physiological; Postoperative Pain; Procedures; Psychometrics; Reader; Research Personnel; Research Project Grants; Risk; Sedation procedure; Signal Transduction; Standardization; Statistical Data Interpretation; Structure; Supervision; Surveys; System; Teaching Hospitals; Testing; Time; Training; Variant; Video Recording; Visual; Vulnerable Populations; Work; algorithm development; base; care burden; care outcomes; care providers; cerebral hemodynamics; college; computer science; convolutional neural network; coping; design; experience; improved; improved outcome; innovation; learning classifier; multimodality; neonatal care; neonate; novel; pain score; patient population; preterm newborn; procedural pain; professor; recurrent neural network; response; side effect; sound; statistics; tool

Most hospitalized infants experience painful procedures as a part of their care. Repeated or prolonged exposure to pain during early life is associated with permanent changes in brain structure and function and may lead to behavioral and developmental disabilities. Caregivers recognize the need to treat pain in infants but may be reluctant to administer analgesics because of potential short- and long-term side effects. Assessing pain in this vulnerable patient population is difficult as their responses are nonspecific and vary based on developmental stage. This difficulty may lead to both over- and under-treatment of pain, placing infants at risk for permanent neurologic injury. We propose to develop models and computer algorithms to automatically and continuously assess pain in neonates based on behavioral and physiological pain indicators. To achieve this goal, the proposal has two aims: aim 1- Continuously and synchronously record contextual, behavioral and physiological signals from neonates in the NICU while they are undergoing prolonged acute pain; aim 2 - developing a multimodal system for assessing postoperative pain in neonates. The proposed system would generate continuous and standardized pain scores comparable to those obtained by conventional nurse-derived pain scores. It would improve care outcomes by enhancing the assessment of pain while decreasing the burden of pain documentation. It can also eliminate issues of inter-rater reliability associated with conventional neonatal pain assessment. The proposed data collection plan was designed based on our preliminary results and statistical analysis. The proposed system will be evaluated using standard performance metrics, such as accuracy, precision, recall, Cohen’s kappa coefficient, or their variations. The interdisciplinary team includes a neonatologist, Dr. Ho, who is a faculty of both USF College of Medicine and College of Nursing. She practices and oversees the clinical research projects at the Tampa General Hospital NICU (the teaching hospital for USF Health). She ensures the team’s access to the NICU patient population for enrollment. She supervises the performance of the study including nursing pain assessment, video recording, and data collection. The team also includes a professor in nursing who specializes in neonatal opioids dependence and pain, an experienced research nurse, two computer science professors, and one statistics professor. The proposed work aligns very well with the NINR’s mission “to promote and improve the health of individuals, families, and communities.” The proposed neonatal pain assessment approach will provide a tool for nurses to assess pain continuously and more accurately. Pain is one of the critical factors affecting the brain development in neonates, especially in preterm neonates, with chronic illnesses. The incidence of perterm neonates exposed to pain in high and pain assessment and management remain a primary focus of neonatal nursing research. For infants in the NICU coping with pain, the proposed nursing research will enable proper pain treatment that will also improve their long-term outcomes and development.

大多数住院婴儿经历痛苦的程序作为他们护理的一部分。重复或长期接触 生命早期的疼痛与大脑结构和功能的永久性变化有关，并可能导致 行为和发育障碍。护理人员认识到需要治疗婴儿疼痛，但可能 由于潜在的短期和长期副作用，患者不愿意使用镇痛药。评估疼痛， 易受伤害的患者人群很难，因为他们的反应是非特异性的，并且根据发育情况而变化 阶段这种困难可能会导致过度和不足的治疗疼痛，使婴儿处于永久性的风险， 神经损伤我们建议开发模型和计算机算法，以自动和连续地 根据行为和生理疼痛指标评估新生儿疼痛。 为了实现这一目标，该提案有两个目标：目标1-连续和同步记录 来自NICU中的新生儿的背景、行为和生理信号，同时他们正在经历长时间的 急性疼痛;目标2 -开发用于评估新生儿术后疼痛的多模式系统。拟议 系统将产生连续的和标准化的疼痛评分，与传统的疼痛评分方法相当。 护士的疼痛评分。它将通过加强疼痛评估来改善护理结果， 减轻疼痛记录的负担。它还可以消除与以下因素相关的评分者间可靠性问题： 传统的新生儿疼痛评估。建议的数据收集计划是根据我们的 初步结果和统计分析。拟议的系统将使用标准性能进行评估 指标，如准确度，精确度，召回率，科恩的卡帕系数，或它们的变化。 这个跨学科的团队包括一位植物学家何博士，他是南佛罗里达大学和 医学和护理学院。她在坦帕的临床研究项目的实践和监督 综合医院新生儿重症监护室（USF Health的教学医院）。她负责确保小组进入新生儿重症监护室 入组的患者人群。她监督研究的表现，包括护理疼痛 评估、录像和数据收集。该团队还包括一位护理学教授， 在新生儿阿片类药物依赖和疼痛方面，一名经验丰富的研究护士，两名计算机科学教授， 还有一个统计学教授拟议的工作与国家核研究所的使命“促进和 改善个人、家庭和社区的健康。”新生儿疼痛评估方法 将为护士提供一个工具，以评估疼痛持续和更准确。疼痛是其中一个关键因素 影响患有慢性疾病的新生儿，特别是早产儿的大脑发育。的 足月新生儿遭受疼痛的发生率很高，疼痛评估和管理仍然是首要任务 新生儿护理研究的重点。对于新生儿重症监护室中应对疼痛的婴儿，拟议的护理研究将 使适当的疼痛治疗，也将改善他们的长期结果和发展。

项目成果

Multimodal neonatal procedural and postoperative pain assessment dataset.

• DOI: 10.1016/j.dib.2021.106796
• 发表时间: 2021-04
• 期刊: Data in brief
• 影响因子: 1.2
• 作者: Salekin MS;Zamzmi G;Hausmann J;Goldgof D;Kasturi R;Kneusel M;Ashmeade T;Ho T;Sun Y
• 通讯作者: Sun Y

Attentional Generative Multimodal Network for Neonatal Postoperative Pain Estimation.

用于新生儿术后疼痛估计的注意力生成多模态网络。

• DOI: 10.1007/978-3-031-16437-8_72
• 发表时间: 2022
• 期刊: Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
• 作者: Salekin,MdSirajus;Zamzmi,Ghada;Goldgof,Dmitry;Mouton,PeterR;Anand,KanwaljeetJS;Ashmeade,Terri;Prescott,Stephanie;Huang,Yangxin;Sun,Yu
• 通讯作者: Sun,Yu

Multimodal spatio-temporal deep learning approach for neonatal postoperative pain assessment.

新生儿术后疼痛评估的多模式时空深度学习方法。

• DOI: 10.1016/j.compbiomed.2020.104150
• 发表时间: 2021-03
• 期刊: Computers in biology and medicine
• 影响因子: 7.7
• 作者: Salekin MS;Zamzmi G;Goldgof D;Kasturi R;Ho T;Sun Y
• 通讯作者: Sun Y