In Tainan, Taiwan’s southern hub, the usually bustling pharmacy division of Chi Mei Medical Heart bustled with activity on this typical afternoon. Pharmacists in crisp white coats move swiftly, accurately filling prescriptions and preparing medication for patients eagerly awaiting their turn at the counter.
Since April, pharmacists have received assistance from a generative AI assistant, or copilot, developed by Microsoft.
Upon clicking the A+ Pharmacy copilot button on the display screen, a patient’s comprehensive medical data is instantly accessible, condensed from various databases onto a single intuitive interface. This streamlined dashboard provides a wealth of information, including treatment lists, surgical details, allergy records, laboratory test results, and a cumulative picture of nursing, medical, and surgical histories – all accompanied by the patient’s unique identifier number, bed assignment, and prognosis.
The One Tab app alerts users to potential adverse medication combinations. A pharmacist can simply click on a photograph of a selected remedy to check if it’s covered by insurance before prescribing the medication.
“According to division head Hui-Chen Su, the design logic mirrors the workflow of pharmacists.”
By leveraging the efficiency of their new copilot system, Su estimates that pharmacists are now able to treat an additional 15 patients daily, effectively increasing capacity by 100% to a total of 30 patients per day. Additionally, this permits pharmacists to allocate more time and attention to patients with complex needs.
Pharmacists at Chi Mei have assigned a collective rating of 4 to 5 stars out of a possible 5, reflecting their overall satisfaction with the product.
Taiwan’s healthcare system is struggling with chronic understaffing, but innovative solutions like generative AI are starting to emerge as a means to support beleaguered medical professionals in their daily work. Artificial intelligence in medical settings has emerged relatively recently, with notable applications in imaging technologies. The new initiative aims to dramatically streamline administrative burdens, encompassing tasks such as extracting data from multiple internal databases, condensing and compiling medical records, and providing patient education materials.
Since November 2023, the 2,500-bed Chi Mei Medical Center has introduced several pilot programs. Medical staff leverage electronic templates to expedite documentation of patient experiences, streamlining the process of compiling admission and progress notes, thereby freeing up valuable time for more critical tasks. Nurses utilize a specialized assistive device as a copilot to facilitate seamless shift changes and mattress repositioning maneuvers. Dietary experts rely on AI-powered tools to offer personalized nutrition advice. By 2019, around 36,000 pilots had been accompanied by co-pilots for some 3,500 individual users.
According to Dr. Hung-Jung Lin, CEO of Chi Mei, the technology not only reduces workload but also ensures patient safety. As healthcare evolves rapidly, our ultimate goal is to provide each medical professional with a dedicated digital companion that streamlines their workflow and enhances patient care.
While it’s still in its infancy, hospitals lack profound understanding of the impact, an initial study of approximately 20 nurses at Chi Mei revealed a significant decline in burnout scores following the introduction of the copilot.
Despite a worldwide shortage of healthcare professionals, Taiwan is particularly affected by this problem.
According to Lin, the biggest challenge in Taiwan’s demographic landscape stems from the combination of an aging population and a relatively low birth rate, which has created a significant shortage of available workers. As the pandemic subsided, the appeal of a medical career dwindled significantly.
Taiwan’s healthcare system is widely regarded as one of the best in the world, featuring a single-payer model similar to those in the UK, where the government plays a central role in providing coverage to its citizens. and Canada. Patients are free to choose any healthcare provider or facility that suits their needs. Taiwanese individuals visit their physicians more frequently, with statistics indicating that they make over twice as many doctor’s appointments per year – approximately 11.4 visits – compared to a global average of around 5.7 visits among the Organisation for Economic Co-operation and Development (OECD) member countries, which comprise mostly high-income nations.
Taiwan lags behind the Organisation for Economic Co-operation and Development (OECD) average in terms of nurse-to-population ratio, with only 7.6 nurses per 1,000 people, compared to the OECD average of 8.9, according to data from Taiwan’s Ministry of Health and Welfare and the OECD.
Healthcare staff are pushed to the limit. Taiwanese hospital doctors typically work over 69 hours a week, exceeding their scheduled 50 hours due to a.
When generative artificial intelligence emerged on the scene last year, hospital administrators swiftly identified opportunities.
Generative artificial intelligence instruments, built upon colossal language models, synthesizing vast amounts of data to produce text, code, images, and more, are regarded as the most significant technological breakthrough since the advent of the internet browser and the smartphone. Currently, Now, , and are pioneering the development of bespoke AI copilot solutions by leveraging their unique domain expertise and practical experience to craft tailored assistants for specific tasks.
Established as a nonprofit medical facility, Chi Mei operates three hospitals under the auspices of the CHIMEI Corporation’s charitable foundation, which also produces plastics and chemical compounds. In a pioneering move, Chi Mei was at the forefront of embracing digital innovation as early as 1995, when it introduced digital medical records.
In 2019, the company established a Clever Healthcare Heart to develop and refine predictive AI models internally. During the 2020 pandemic, for instance, the centre developed predictive models to forecast which COVID-19 patients would require ventilation support based on vital signs, as well as how quickly a patient could be weaned off mechanical ventilation, according to Chung-Feng Liu, the centre’s director and professor of information management.
By 2022, the Clever Healthcare Heart leveraged Microsoft 365 and Energy BI to streamline administrative processes and analyze hospital data, including benchmarking quality metrics? When Microsoft’s Azure OpenAI Service debuted last year, it integrated the hospital’s extensive datasets with the generative AI platform, ensuring the secure storage and processing of sensitive patient data while adhering to the hospital’s confidential protocols.
In November 2023, Chi Mei unveiled a series of copilot models, each bearing the distinctive ‘A+’ moniker. Moreover, a team of A+-credentialed professionals, including physicians, nurses, pharmacists, and nutritionists, works alongside an A+ Patient Safety copilot that proactively identifies patients at risk of falls by monitoring keywords such as “bed sores” or “low blood pressure,” and provides personalized safety recommendations. Elevating the Standard: An Advanced National Assessment Program Facilitating Collaborative Learning for Medical Professionals to Stay Current with Ongoing Education Requirements.
According to Dr. Chia-Te Liao, a cardiologist and director of Chi Mei’s Center for Evidence-Based Medicine and Health Policy, medical professionals previously dedicating an hour to writing reports can now generate and edit comprehensive reports based on their notes within 15 minutes, simply by clicking “affirm”.
Cockpit personnel can create tailored training materials for individuals grappling with diverse conditions, rather than distributing generic pamphlets on each condition.
A healthcare technology innovation team, comprising Liao representing medical doctors and Liu representing technologists, collaborated to gain consensus by organizing workshops where medical doctors, nurses, and other stakeholders contributed to designing copilot tools tailored to their workflows.
There was some early resistance. Some medical doctors were quoted as saying: “We don’t need AI.” I’m a physician.”
Initially, nurses harboured concerns that the introduction of a copilot role might render their expertise obsolete. Liao emphasized that hospital leaders must convene an assembly to address and alleviate the concerns of those in their care.
The CEO clarified that the purpose of the copilot program was to provide support to nurses, not replace them. However, we sincerely hope that you will be able to depart the hospital on schedule, thereby freeing up valuable time for yourself and your loved ones, Liao emphasized.
Currently, approximately two-thirds (63%) of Chi Mei’s 95 pharmacists utilize a copilot, while nearly half (47%) of the 2,000 registered nurses and roughly one-third (23%) of the total 700 physicians are also using this assistance mechanism. Two-thirds of the 20 nutritionists working at the middle have also taken up using a copilot.
While all copilots are built on the Azure OpenAI Service platform, their GPT-based models, or Large Language Models (LLMs), vary depending on the specific requirements. While some residents have adopted GPT-3.5, others rely on the more advanced GPT-4 for analyzing complex medical data.
Nursing staff have significantly reduced the time required for documenting patient information, including physician orders, admission summaries, test results, and other essential details, from a range of 10-20 minutes to under 5 minutes.
According to Yu-Chen Tung, a nurse at Chi Mei, this efficiency enables the team to devote more premium attention to patients by their bedside.
Chi Mei is actively improving its copilot capabilities, with a primary focus on linguistic advancements. At times, it inappropriately employs a non-medical professional’s timeframe rather than a medical practitioner’s. Nurses have consistently evaluated the copilot positively, awarding it a rating of three to four out of five stars, according to Tung.
According to Liao, a major criticism of our service until now has been the quality of language used.
Coproats often insert a simplified Chinese character as a substitute for the standard script employed here. Despite Liao’s reservations, another critique is that the anecdotes appear eerily similar to having been crafted by artificial intelligence algorithms.
Despite being a general observation, Chi Mei’s CEO, Lin, noted that the company’s experience with AI had exceeded their initial expectations, “we found that AI performed beyond what we anticipated.”
By integrating disparate data sources into a unified platform, encompassing not only medical information but also socioeconomic factors gleaned from social worker encounters, and “insightful details that medical practitioners often overlook,” generative AI, as Lin emphasized, serves as a poignant reminder for healthcare professionals to prioritize empathy in their practice.
