Practical exploration of artificial intelligence in the field of medical nursing care

Sun Yan-guo, Cui Di

Information Center of the Second Affiliated Hospital of Dalian Medical University, Dalian 116023,Liaoning Province, China

Abstract In the current era of continuous development of artificial intelligence (AI) technology, medical care has become an important research and application field in the AI industry. How to achieve an effective application of AI technology in clinical nursing field is currently a key research topic in the industry. The authors have discussed and introduced the architecture, functions and practical effects of the AI intelligent nursing system based on AI and mobile computing technology,constructed by the Second Affiliated Hospital of Dalian Medical University. This system has been officially put into operation in the Second Affiliated Hospital of Dalian Medical University since November 2018. A complete set of clinical nursing corpus has been established, equipped with customized smart hardware PDA products and integrated with digitalized smart ward system,which can improve overall quality of clinical nursing care and management, subvert the traditional information way, and promote the establishment of smart digitalized ward system. A novel model of humanized and refined ward management that is beneficial to medical treatment and scientific research has been constructed.

Key words: artificial intelligence; speech recognition; natural language processing; clinical care

Introduction

On April 13, 2018, the National Health Commission of the People’s Republic of China has issued the “Notice on the Issuance of National Hospital Information Construction Standards and Specifications (Trial)”. Artificial Intelligence(AI), Big Data, Internet of Things and Cloud Computation appear as emerging technologies in the top level for the information construction of tertiary hospitals. The introduction of AI intelligent nursing system in our hospital is the first innovative practice to truly apply speech recognition technology and natural language analysis technology on clinical nursing in China, providing new application examples for mobile medicine, intelligent medicine and other work scenarios, forming an innovative application system for intelligent speech recognition[1].The main principle is to provide nurses with an intelligent terminal tool through intelligent hardware and software combination, so that nurses can get rid of the restrictions by different application systems and by the place of use.Using simple voice commands, nurses are able to complete 80% of their daily work, thus realizing the intelligent,mobilized and simplified nursing work.

Existing Problems of Informatization in Clinical Nursing Work

Various information systems, cumbersome operations and increased workload

During the present clinical nursing work, the construction of basic information system has been gradually improved.With the requirements of refined management, superior management and policy system for medical practice,an increasing number of clinical auxiliary systems are introduced to performed standardized management of nursing behaviors from all aspects of clinical work,accordingly a large amount of information needs to be collected, summarized, and analyzed. Therefore the workload of nurses rises rather than falls, resulting in insufficient available time to provide better nursing service.Nurses have to operate different information systems every day to complete their daily work, and are often busy with system login, switching, and function search. Clinical nursing practice turns to be managed in more refined manner and covers various systems and modules, leading to cumbersome operations and relatively low efficiency.It requires a lot of manual operations by nurses, which not only wastes effective working time of nurses, but also cannot realize the patient-cored comprehensive medical care, improve the quality of nursing service, prevent medical risks, or improve the doctor-patient relationship.

Delayed information acquisition and predominant repetitive work

According to the feedback from the statistical data in hospitals, about 30%-45% of the nursing work is currently occupied by repetitive work every day, and it requires to consult a large amount of information during the nursing service process to carry out nursing work and answer patients’ questions. Meanwhile, it is also necessary to record a large amount of patient’s clinical data during the nursing service process, so that the physicians can have a comprehensive understanding of the patient’s overall condition. The existing method for information recording and query are mainly based on desktop computer records, and PDA bedside records are used as auxiliary methods. PDA nursing care only integrates part of the bedside functions in the HIS system, and cannot fundamentally solve the core needs of informatization extending to the bedside. When a nurse needs to obtain certain information, he/she often has to go back to the computer in the nurse station and check information. This work is repeated every day and occupies a lot of nurses’working time. At the same time, omissions and negligence are prone to occur in the process of information transfer,causing omissions accident or medical disputes. How to reduce the burden of daily nursing work and repetitive labor and improve the quality of nursing care by using intelligent technology, has become a very prominent problem.

Insufficient hardware resource and failure of smart advantage performance by smart devices

As the concept of hospital informatization construction often focuses on the doctor-cored ward informatization construction, the core resources basically give priority to doctor-terminal construction, resulting in less investment in clinical nursing care. At the same time, due to the space limitation in nurse stations, the general nurse station is only equipped with 2-4 computers. Most nurses have to share computers for operations. In addition, there are still many nursing works that are still managed by written records, and the nurses are often unable to truly obtain work convenience due to informatization. In recent years,with the development of smart PDA devices, increasing hospitals have applied mobile nursing systems to realize mobile nursing work and extend information services to patients’ bedside. However, most of the current mobile nursing systems transfer traditional HIS functions to mobile nursing systems through only transforming the information system carrier from a computer to a PDA.There has been no change in the system use mode, and no transformative innovations have been initiated in business development. Therefore, mobile nursing system has gradually become a performance appraisal tool used by the head nurses for the management of nurses, rather than a closed-loop management tool for the entire nursing process expected at design.

Significance of AI Intelligent Nursing System in Clinical Nursing Practice

Changing traditional working methods and improving work efficiency

Nurses have no longer to log in the information system as before, or perform complicated operations such as selection, switching, clicking and logging-in. They only tell the PDA what they want to do. Then AI intelligent nursing system can intelligently understand the nurse’s voice commands, automatically complete related operations, and give a feedback upon the operation results. For example,“Query all patients who currently need a measurement of their blood pressure”, “Record the temperature of patient in bed 3 as 36.8℃”, “Which patients currently need a preoperative health education”, “Are there currently unexecuted medical orders”, “Record pre-lunch blood glucose value of patient in bed 5 as 10 mmoI/I” and so on. The system can cover all clinical nursing works. AI intelligent nursing system is able to really understand the meaning of nurses as long as they say what they need to do at any time and any place. Through natural language processing, the system then performs an accurate analysis of the recognition results to define “who is talking”,“what to do”, “how to do”, “expectation results”, accurately understands the nurse’s intentions, and automatically completes related tasks. This can effectively reduce the workload of medical staff, improve work efficiency, enable them to quickly respond to patient’s requirements, and improve service quality.

Presenting aggregated information and achieving one-stop intelligent service

Nurses have no longer to use multiple systems or switch between various systems. Only using the AI intelligent nursing system, nurses are able to solve the existing clinical business. AI intelligent nursing system can customize the business model, business process and business data,comprehensively integrates with the basic data information in hospitals, and produces an interconnection with the HIS system, LIS system, PACS system, EMR system, CPW system, surgical anesthesia system and other in-hospital systems through the integrated platform. The system can aggregate and display patient’s information data, medical order information data, examination information data,laboratory information data, medical record information data, cost information data and surgical information data according to different business, and present customized information content according to business needs, while performing multi-system data synchronization for semantics, and ensuring no difference in data between systems and real-time synchronization.

Getting rid of the restrictions by the place of use and realizing comprehensive mobile office

By using medical Internet of Things technology, intelligent perception technology, information integration technology and intelligent decision-making technology, AI intelligent nursing system quickly promotes the transformation of medical models[2], and truly realizes the networking,mobilized and dynamic medical nursing work. Nurses have no longer to complete related tasks in the computer at the nurse station as before. They get rid of the restrictions by the place of use. Under the coverage of the network,they are allowed to complete nursing-related information work through the use of smart terminal PDA in any place such as wards, ward corridors, offices, dispensing rooms and so on.

Constructing AI intelligent voice-open service platform and providing professional intelligent voice services in the nursing field

The core of the AI intelligent nursing system is to construct an AI medical energizing open-access system that can be widely used in the medical industry, and provide professional medical nursing voice recognition services, semantic interpretation open-access capability services, and visualization analysis services of multidimensional isomerizing data models in wards, for other information systems in the hospitals, thus forming an open-access, intelligent and automated speech recognition and semantic interpretation service platform. In addition,it establishes standard model references for semantic recognition and business automation in the medical field,and constructs open standards for medical energizing.This system can comprehensively solve the relevant problems, such as lack of AI landing demonstration in the field of medical ward services, long R&D cycle, high implementation risk, uneven capability levels and repeated constructions, it also promotes the industrialization and application of the system, as well as the integration and collaborative innovation of production, learning,research and application, impels the construction of smart hospitals, and effectively drives the development,exploration and innovation of smart medical services in hospitals. A refined, comprehensive and efficient AI intelligent nursing system is established, to realize the digitization, integration and intelligence of medical services such as patient diagnosis and treatment and ward management, and promote the establishment of a digital intelligent medical ward system.

Main Implementation Approach of AI Intelligent Nursing System

The system is equipped with complete business processing architecture, voice converters, language interpreters,dialogue flow controllers, task controllers and output system, these components work together to establish a clinical corpus based on medical nursing care, make full use of speech recognition technology and natural language analysis technology, summarize a large number of clinical nursing text data, integrate the established software and hardware platforms of medical big data, conduct largescale training by using cloud computation, big data and machine learning technologies, and then perform speech analysis on the speech text recognized, build a trainable NLP pipeline. Based on deep-learning sentence segmentation, specific part-of-speech tagging in medical industry and self-correction algorithms, it customizes medical voice recognition models that is applicable to medical nursing scenarios, so that the system can fully understand the meaning of nurse’s voice commands, and precisely recognize medical terms, specific environmental languages and local special words. After the semantic understanding is completed, the system will operate according to the nurse’s semantics, and identify medical professional terms with high accuracy so as to truly understand the semantics of medical staff.

Conclusion

In the near future, AI speech recognition and natural language processing technology will play an important role in the medical field, which will subvert the work and service mode in the wards, change the working habits of nursing staff, efficiently promote the realization of highquality nursing care in the wards, and even subvert the medical model. For the hospital management, this system can scientifically and effectively implement high-quality medical services, promote intelligent management in the wards, establish dynamic and innovative service concepts,and create value.