Improving communication outcomes in children with cleft palate in rural India

项目摘要/摘要 儿童唇腭裂(CLP)的治疗包括一次手术修复唇裂， 然后进行二次治疗，如咽瓣手术和语音治疗，目的是 提高言语效果和生活质量。印度每年出生的儿童数量位居第二 与中电并驾齐驱，仅次于中国。在像印度这样的中低收入国家，大多数 患有CLP的儿童生活在农村地区，在最初的唇裂手术损伤后仍未得到治疗。 印度医学研究委员会(ICMR)进行的一项研究报告称，50%-55%的患有 中电要求进行后续手术以改善语言能力，但这些儿童中的绝大多数没有得到应有的治疗 获得临床服务的机会有限。在印度大多数农村CLP病例中，干预措施在小学后结束 手术和患者及其家人接受作为这种障碍的结果的糟糕的语言结果。 这是不幸的，因为许多情况下，这些言语问题是可以通过二次手术或 言语治疗；然而，确定谁将受益于这些的资源和诀窍 干预措施集中在城市地区，而农村地区的干预措施不知道额外的治疗 是有可能的。为了确定是否有必要进行二次手术或语言治疗，临床医生会听取 儿童讲话中的过度鼻音。这是演讲中的一个微妙的暗示，表明 存在可矫正的腭咽闭合功能障碍，导致语音清晰度降低。这是一个高度的 乡村诊所的社区医护人员没有接受过培训的专业技能。结果, 农村地区的儿童得不到有可能改善的次级干预措施的好处 言语结果和生活质量。 我们提出了一种基于人工智能(AI)的移动健康(MHealth)应用程序，用于 确定哪些患有CLP的儿童将从改善沟通的干预中受益 在印度农村地区取得的成果。这是一个多学科的合作项目，由亚利桑那州立大学 印度达瓦德理工学院和全印度语言与听力学院。我们 将使用印度主要语言卡纳达语的CLP儿童现有语音数据库，以 为印度语言开发一个超鼻音预测算法；我们将在 一个mHealth应用程序(R21：SA1)。我们将与AIISH合作对该应用程序进行初步临床评估，并 算法(R21：SA2)，然后基于此评估(R33：SA3)改进APP和算法。最后， 我们将向社区卫生工作者分发mHealth应用程序，以便在乡村诊所进行评估(SA4和 SA5)。应用程序开发的里程碑和应用程序的初步验证和可用性将实现 以便从R21移动到R33。

Project Summary / Abstract The treatment of children with cleft lip and palate (CLP) involves the primary surgical repairment of the cleft, followed by secondary treatments like pharyngeal flap surgery and speech therapy for the purpose of improving speech outcomes and quality of life. India has the second highest number of children born yearly with CLP in the world, second only to China. In lower- and middle-income countries like India, most of the children with CLP live in rural areas and remain untreated after the primary surgical impairment of the cleft. A study conducted from Indian Council of Medical Research (ICMR), reported that 50-55% of children with CLP require follow up surgery to improve speech, but the vast majority of these children are not treated due to limited access to clinical services. In most rural CLP cases in India, intervention ends post-primary surgery and patients and their families accept poor speech outcomes as a consequence of the disorder. This is unfortunate as, many times, these speech problems are correctable via the secondary surgery or speech therapy; however, the resources and know-how for determining who would benefit from these interventions are clustered in urban areas and those in rural regions are unaware that additional treatment is possible. To determine if secondary surgery or speech therapy are necessary, clinicians listen for signs of hypernasality in the speech produced by children. This is a subtle cue in the speech that indicates the presence of correctible velopharyngeal dysfunction resulting in reduced speech intelligibility. This is a highly specialized skill that community healthcare workers in rural clinics have not been trained in. As a result, children in rural areas go without the benefit of secondary interventions that have the potential to improve speech outcomes and quality of life. We propose an artificial intelligence (AI)-based mobile health (mHealth) application for the identification of children with CLP who would benefit from interventions to improve communication outcomes in the rural areas of India. This is a multidisciplinary, collaborative project between Arizona State University, the Indian Institute of Technology Dharwad, and All India Institute of Speech and Hearing. We will use an existing database of speech from children with CLP in Kannada, a primary language in India, to develop a hypernasality prediction algorithm for the Indian language; we will integrate this algorithm within an mHealth app (R21: SA1). We will work with AIISH to perform an initial in-clinic evaluation of the app and the algorithms (R21: SA2), then refine the app and algorithms based on this evaluation (R33: SA3). Finally, we will disseminate the mHealth app to community health workers for evaluation in rural clinics (SA4 and SA5). Milestones for development of the app and initial validation and usabililty of the app will be achieved in order to move from the R21 to the R33.