封閉制度下之創新擴散：AI在醫療應用之個案研究

為了解決全球老化以及實現個人化醫療的願景，近十年以AI導入醫療照護產業的「智慧醫療」又重新捲土重來，尤其2020年新冠肺炎影響，更是將AI導入醫療領域的必要性更往前推進，AI導入藥物探索在這兩年有著顯著的躍進，科學家們迫切希望可以雪恥過去的失敗經驗。過去藥物的研發到上市平均需要10年的光陰，AI應用最大的效益是在自動化上游資料處理過程中重複性高的流程，精準配對、減少人為錯誤並縮短不必要研發時間，增加全新開發候選藥物數目、減少藥物發展後期失敗率。在這個急需解決全球公共衛生事件的現在，是個最值得去開發跟投資的一塊領域了之一了。藉由AI的機器學習與深度學習技術，已被相信能有效地提高醫療照護流程中的許多層面的效率，AI也因其能同時處理不同種類且龐大的資料量，其所提供的決策建議被認為具有更高的精準性，在照護實務中則可直接有助於縮短以往根據經驗法則給予處置的錯誤期，減少了不必要的醫療開銷。與智慧醫療相關的議題討論相當多，無論是從產業發展角度或者是以醫師等醫療服務提供者的角度去探討，多數對於AI在醫療上的發展抱持著厚望，尤其現在有相關法規、政策的支持，AI的導入勢必會為醫療產業開創一個新的局面。本研究選擇AI在醫療產業應用作為背景，試圖從近年來AI在醫療這個封閉領域中的成果、各大科技廠的跨域競爭以及醫院端又如何將這些技術導入，讓終端使用者(醫護人員)接受等面向，來分析未來AI在醫療應用上的發展及挑戰。對於智慧醫療的發展之下，企業們該如何在此封閉的系統下與醫藥界的專家合作? 又或者這些科技大廠在跨域競爭所面臨的挑戰該如何解決? 研究的方式主要透過本論文AI導入醫療產業中，代表不同面向的四個個案，從其研發成果、跨域挑戰的利基，以及法規的限制進行探討，試圖回答本研究所提出的問題，並在最後的研究與建議中，歸納出相關的管理意涵。

關鍵字

AI ；智慧醫療；精準化醫療；深度學習；跨域競爭；制度理論；破壞式創新

並列摘要

In view of the need to address the two visions of global aging and the realization of personalized medicine, "smart medicine", which has been restarted and introduced into the medical care industry with AI in the past ten years. Especially affected by COVID19 in 2020, the need to introduce AI into the medical field is even more valued and promoted. The introduction of AI into drug discovery has made significant progress in the past two years. Scientists are eager to use successful results to reverse past failures. In the past, it took an average of 10 years for a drug from development process to launch in the market. The biggest benefit of AI application is the highly repetitive process in the automated upstream data processing steps, which can accurately match, reduce human error and shorten unnecessary development time; thereby increasing the number of newly developed drug pipeline and reducing the failure rate of drug development. The real society is an era in which global public health incidents are urgently needed to be resolved. The above is one of the most worthwhile areas for development and investment. The machine learning and deep learning technology possessed by AI has been believed to be capable of effectively improving the efficiency of many aspects of the medical care process. In addition, because AI can process different types and huge amounts of data at the same time, the decision-making suggestions it provides are considered to provide higher accuracy. In nursing practice, AI applications can directly help shorten the error period that humans have dealt with based on experience in the past, thereby reducing unnecessary medical expenses. At present, there are quite a lot of discussions on topics related to smart medical care. Whether from the perspective of industrial development or from the perspective of medical service providers such as doctors, most of them hold hope for the development of AI in medical care. In particular, with the support of relevant regulations and policies, the introduction of AI is bound to create a new stage for the medical industry. This research focuses on the application of AI in the medical industry as a background, trying to analyze from the achievements of AI in the closed field of medical care in recent years, the cross-domain competition of major technology factories, and how hospitals can introduce these technologies for end users (medical personnel) to accept, furthermore, to discuss the future development and challenges of AI in medical applications. Regarding the development of smart medical care, how should companies cooperate with experts in the pharmaceutical industry under this closed system? And how should the various challenges faced by these big technology companies in cross-domain competition be solved? The method of this research is mainly through the introduction of AI into four cases represented by different aspects in the medical industry, discussing its research and development results, the niche of cross-domain challenges, and the restrictions of laws and regulations, trying to answer the questions raised by this research. Finally, in the research and recommendations, this research summarizes the relevant management implications.