Novo Nordisk and Machine Learning
Technology has been continuously evolved, and new resources and systems more and more intelligent come emerging throughout the years. Natural language processing (NLP) principles and Machine Learning have contributed actively to the scientific advances, and Novo Nordisk wants to explore more these new technologies of intelligent systems, to boost its clinical researches at the health field once that Real-world data (RWD) has not been being satisfactory to the corporation by itself. The Novo Nordisk is willing to generate access improved to its data silos, data lakes, migrating the NLP platform and Tableau, integrated to the Amazon Web Services (AWS) cloud-based global big data platform. However, there are ethical concerns about how these researches and clinical trials are conducted.
BACKGROUND
Novo Nordisk is a big successful pharmaceutical organisation, which is more commonly known for its set of products at the diabetes treatment, where the company is a market leader (Bakiny, 2019). Novo Nordisk is a multinational healthcare organisation with over 85 years of excellence, innovation and expertise in diabetes care. The corporation also is a leader in haemophilia, developing hormone therapy and treatment. Novo Nordisk has its matrix located in Denmark, the company has over 30,000 employees in 76 countries and markets its products in over 170 countries (Novo Nordisk History, n.a).
Intelligent systems are machines advanced in technological terms, which perceive and react to the environment around them. Intelligent systems may assume several types, from automatic vacuums such as Roomba to face recognition applications to customised Amazon shopping tips (University of Nevada, n.d). These intelligent systems are revolutionising several industries, such as health care, logistics, automation, education, entertainment, security, military application, manufacturing, robotics and transportation. The systems assist in enhancing the energy efficiency, flexibility and quality of those systems. Intelligent systems are dynamic and use a wide variety of innovations technology such as cybersecurity, AI-artificial intelligence, NLP — natural language processing, machine learning, embedded CPUs, wireless networking and visual signalling and distributed storage (Lewis University, n.d).
Novo Nordisk initiated its researches working with RWD, which is information data, derived from sources other than conventional clinical trials that currently has been becoming more and more relevant in decision making in the healthcare industry. The growing ubiquity of biometric monitoring systems in the form of mobile phones, smartwatches and other wearable electronics is one of the most promising features of the RWD studies. These smart devices that have become instruments to gather quantitative biological data such as pulse rate, blood pressure, electrocardiography, mobility and sleep cycles have the potential to improve massively clinical trials due to powerful capacity. In traditional RCTs, these parameters are registered at discrete time points in a regulated environment. The use of biometric tracking systems could permit these criteria to be continuously gathered and remotely, turning the data even more expressive in terms of accuracy and efficiency at the therapies (Nuventra, 2019).
RWD has demanded much effort in terms of manual scanning and extraction techniques to go over and process severals medical sources. The RWD extraction process took work-intensive and ineffective. The proceed database is difficult to search and non-interactive, and as data quantity and requests for accessing raised the scalability significantly became a problem (IQVIA 2020, p. 02). For these reasons, Novo Nordisk sought new methods and tools such as NLP, machine learning and Tableau as a solution to those problems.
METHODS AND TOOLS
Methods Applied: Systems Integration — In the first part, the team of Novo Nordisk has iteratively filtered, enhancing the application algorithms to extract well-structured clinical information from non-structured RWD sources.
The second part of the project is to migrate all the applications to the cloud AWS solution base on Big Data and OASIS. Python and Zeppelin Notebooks were used to make a data pipeline, extracting RWD sources to the data lake, running in the NLP application, and deposits the extracted outcomes back into the data lake automatically for evaluation, visualisation and manipulation by Tableau (IQVIA 2020, p. 03).
Figure 1. Linguamatics NLP Extracting from RWD and other Sources (Reed & Breyette, n.d).
Natural Language Processing (NLP) is a linguistics sub-area, computing science and AI artificial intelligence referred to cooperation human language to the computer and interactions between them, and how computers can be programmed to handle and interpret vast quantities of natural language data. NLP text-mining platform Linguamatics makes effective, produce valuable perspectives and reachable around the business.
Machine Learning is an implementation method of artificial intelligence AI, which generates the systems the ability to learn and build on knowledge without being directly programmed automatically. Machine learning has as its objective the creation of computer programming that may access data and learn by themselves (Luxton, 2016).
Tableau is a data analytics application, which is commonly used for BI, business intelligence. This visualisation tool generates dynamic diagrams, graphs and mapping in the form of dashboards and worksheets to achieve business perspective (Kaur, 2017). For information about tools such as Big Data and AWS consult the appendices.
Figure 2. Data Pipeline (Reed & Breyette, n.d).
RESULTS
The adoption of intelligent systems and tools speeded up Novo Nordisk’s researches and consequently maximised its business success in many important areas and aspects. It is what did Novo Nordisk win the company’s Innovative Practices Award in 2018 (Moore, 2020).
· With the adoption of new intelligent systems and integration among them, Novo Nordisk has saved around $100,00 annually. Previously, tasks that used to take six weeks to finish through a high number of employee now is done in a few hours.
· Within Health Economics Results Research (HEOR) ethnographic insight research, the NLP framework is used to mine transcripts of patients, caregiver consultative boards and customer committees and to extract useful information from over hundreds of pages of conversational content.
· In the continuing analysis of clinical trial protocol anomalies, the NLP platform replaced the unreliable, quarterly, manual method with extracted data made accessible via dynamic dashboards which generated deeper understanding and perspectives.
· Within social media project to recognise prominent obese opinion leaders, the NLP system was applied to mine Twitter feeds. Recently, medical affairs teams used to pay dozens of thousands of dollars to outsource this task, however, the NLP application allows Novo Nordisk company to reuse these capabilities in-house.
· Novo Nordisk will be able to work on many more different projects at the same time more effectively and efficiently. The organisation is using NLP to find out ambitious and competitive perceptions from Dow Jones DNA news data (IQVIA 2020, p. 04).
DISCUSSION
The health care sector is hugely relevant, thus the AI and its resources have been revolutionised positively, the way and precision on the health researches development, treatments and diagnostics. However, there are significant ethical aspects that involve the use of NLP and intelligent systems overall.
The innovation systems speeded up the health care sector in terms of technology, efficiency and effectiveness, for example, diagnosing skin cancer much more accurately, better than a dermatologist specialist with the highest medicine degree and faster than a decade of too extensive and expansive medical education. (Rigby, 2019). Despite benefits, also there are constant concerns about the Health field regard to ethics and social issues. Some of them:
· Transparency and Responsibility
· Data Bias, justice, and Equity
· Effects on Patients
· Data Security and Privacy
· Malicious use of AI
· Challenges for Governance
· Legislation in terms of Breaking ethics using AI.
Figure 3. Guidelines for Trustworthy AI issued by European Commission, 2018 (Muthuswamy, 2019).
Clinical test Ethics Novo Nordisk endorsed clinical trials are undertaken using a single global standard. This standard is established in compliance with the Helsinki Declaration and the Recommendations on Good Clinical Practice (GCP) of the International Conference on Harmonization (ICH). The Helsinki Declaration, established by the World medical committee, is a reference pattern of ethical guidelines for clinical studies. The ICH-GCP Guidelines aim to ensure the rights and preservation of trial participants and maintain the legitimacy of the results of the tests. For those who have been engaged in clinical studies, the concerns of the patients of the study must take precedence over the concerns of science (The Blueprint for Change Programme Novo Nordisk, n.d).
RECOMMENDATIONS
· Follow the successful leading corporations, which have used the same systems and learn with them mistakes and hits.
Roche’s methodology is an example of patient-centred medicine production (FDA), which means that clinical illness models environ patient needs. The organisation Linguamatics congratulated Roche on winning the well-deserved award and on their creative use of technology to advance scientific research. 2019, is the third straight year that Linguamatics or one of Linguamatics’ associated has received the Bio-IT World Award (Moore, 2020) Roche also is an Australian health care successful company specialised in diabetes studies (Roche, n.d).
· That’s important to know what the competitors are doing, analysing their studies, researches, surveys, results, growth, finding out how they are applying the technological resources and intelligent systems, and then to learn with them.
· It is fundamental to keep studying and exploring continuously the new technologies, which are new intelligent systems, resources and tools, to maintain the company success growth.
· Outside of the Health field, there are other examples from the big corporations, such as Google, Netflix, Amazon and Uber, which are from different industries that also can be observed, followed, studied, adapt and applied in the health sector. Explore that.
· The ethical aspects are fundamental to preserve the patient’s data integrity, treatment and interests above the researches.
The patients’ data cannot be shared or disclosed without previous authorisation from them. The Health Companies Researches must obtain this authorisation by contract with the patient allowing the disclosing of their details, data and results within the research. Otherwise, the disclosing of patient information without previous authorisation may be considered a crime.
CONCLUSION
The integration of new intelligent systems and Novo Nordisk’s implementation text mining tools, together with Tableau Visualisations and AWS cloud-based pipeline, allows efficient and systematic analysis across six valuable sources of data integrating real-world evidence and scientific research. Through migration to the Aws cloud, the data pipeline saves considerable time and ensure stakeholders with on-demand insights to support evidence-based decision-making. Novo Nordisk is expanding usage of cloud technology for big data efficiency, connectivity, versatile provisioning, and robust computational and data analytics tools, which is not a trend exclusive within the pharmaceutical industry, but rather in the many other sectors. If Novo Nordisk keeps going to respect the ethical issues and prioritise the patient interests beyond its researches, the corporation will definitely consolidate itself in the market and achieve its goals successfully.
APPENDICES
Figure 4. Tableau Dashboard (Kaur, 2017).
Figure 5. Linguamatics Award-Winning NLP Platform (Linguamatics, n.d).
Figure 6. NLP Interaction Context (Boulwafa, 2020).
Figure 7. Natural Language Processing (NLP) — Applications of NLP (Boulwafa, 2020).
Figure 8. The Suggestion for a process to “Ethics by Design” in an Organization (Leidner & Plachouras, n.d).
Figure 9. Oasis (Reed & Breyette, n.d).
Figure 10. Amazon Web Services (AWS) (Amazon, n.d).
Figure 11. Big Data Definition (Segal, 2019).
REFERENCES
Amazon. (n.d). Cloud computing with AWS. Retrieved from https://aws.amazon.com/what-is-aws/
Bakiny, J. (2019, December 31). Up More Than 20% This Year, Is Novo Nordisk Worth Buying? Retrieved from https://www.fool.com/investing/2019/12/31/up-more-than-20-this-year-is-novo-nordisk-worth-bu.aspx
Boulwafa, M. (2020, April 11). Natural Language Processing (NLP) — What is NLP ? Retrieved from https://medium.com/analytics-vidhya/natural-language-processing-nlp-introduction-fe48e9b7ec8d
IQVIA. (2020). Using Natural Language Processing At Novo Nordisk To Generate Actionable Insights From Real World Data. Novo Nordisk Case Study
Kaur, P. (2017, July 27). Tableau for Beginners — Data Visualisation made easy. Retrieved from https://www.analyticsvidhya.com/blog/2017/07/data-visualisation-made-easy/
Leidner, J., & Plachouras, V. (n.d). Ethical by Design: Ethics Best Practices for Natural Language Processing. Retrieved from https://www.aclweb.org/anthology/W17-1604.pdf
Lewis University. (n.d). What are Intelligent Systems? Retrieved from https://online.lewisu.edu/mscs/resources/what-are-intelligent-systems
Linguamatics. (n.d). Linguamatics NLP Platform Overview. Retrieved from https://www.linguamatics.com/products/linguamatics-natural-language-processing-platform
Luxton, D. D. (2016). An Introduction to Artificial Intelligence in Behavioral and Mental Health Care. Retrieved from https://www.sciencedirect.com/topics/psychology/machine-learning
Moore, M. A. (2020, July 2020). Roche awarded the 2020 Bio-IT World Innovative Practices Award for use of IQVIA Linguamatics NLP in patient-focused research. Retrieved from https://www.linguamatics.com/blog/roche-awarded-2020-bio-it-world-innovative-practices-award-use-iqvia-linguamatics-nlp-patient
Muthuswamy, D. (2019). Ethical issues in Artificial Intelligence in Healthcare. Retrieved from https://www.itu.int/en/ITU-T/Workshops-and-Seminars/ai4h/201911/Documents/S4_Vasantha_Muthuswamy_Presentation.pdf
Novo Nordisk History. (n.a). Retrieved from https://www.novonordisk.com/content/dam/Denmark/HQ/aboutus/documents/HistoryBook_UK.pdf
Nuventra. (2019, December 5). What are Real-World Data and How Can they Benefit Drug Development? Retrieved from https://www.nuventra.com/resources/blog/real-world-data/#:~:text=Real%2Dworld%20data%20(RWD),avenue%20for%20achieving%20product%20approval.
Reed, J., & Breyette, T. (n.d). Using Natural Language Processing to transform real world data. Retrieved from https://www.lexjansen.com/phuse/2019/rw/RW05.pdf
Rigby, M. J. (2019). Ethical Dimensions of Using Artificial Intelligence in Health Care. Retrieved from https://journalofethics.ama-assn.org/article/ethical-dimensions-using-artificial-intelligence-health-care/2019-02
Roche. (n.d). Roche. Retrieved from https://www.roche-australia.com/
Segal, T. (2019, July 5). Big Data. Retrieved from https://www.investopedia.com/terms/b/big-data.asp
The Blueprint for Change Programme Novo Nordisk. (n.d). Retrieved from https://www.novonordisk.com/content/dam/Denmark/HQ/sustainablebusiness/performance-on-tbl/more-about-how-we-work/Creating%20shared%20value/PDF/blueprint-diabetes-clinical-research.pdf
University of Nevada. (n.d). What are intelligent systems? Retrieved from https://www.unr.edu/cse/undergraduates/prospective-students/what-are-intelligent-systems#:~:text=Intelligent%20systems%20are%20technologically%20advanced,to%20the%20world%20around%20them.&text=The%20field%20of%20intelligent%20systems,dynamic%20physical%20
