數位化｜掌握數據，創造更多價值：用大數據提供更好的服務 More data for more users means more value

Ferran Urgeles，百靈佳殷格翰Dataland計畫經理（Program Manager of the dataland initiative at Boehringer Ingelheim）。

無論是在線上還是在眾多虛擬世界的領域中，演算法已成為日常生活的一部分。不妨以現代工業過程各階段所產生和記錄的數據為例，無論是在生產、溝通、與客戶接觸、會計還是研究方面，都會產生大量的數據。涉及的資訊量過於龐大，人類的大腦難以處理，但演算法能分析這些數據，並提供新的見解，發現模式並建立連結。

「科技讓我們更了解哪些過程對公司的績效產生了影響，以及如何產生的影響。」百靈佳殷格翰計畫經理Ferran Urgeles表示：「這將使我們發揮Dataland的全部潛力，帶給我們前所未見的新見解。」

Urgeles先生說：「百靈佳殷格翰的最終目標是當然是幫助我們更快速、更可靠地開發藥物。」

一個安全、可存取的數據生態系統

Dataland生態系統以一個數位平台為核心，整合公司各領域的數據，並以易於理解的方式，供模擬和數據分析等用途使用。

「在Dataland生態系統中的所有數據都會經過數據安全性和合規檢查，並使用最先進的安全措施來保護。」Urgeles先生說：「這個生態系統從2022年底開始在百靈佳殷格翰使用，應用在實際案例上。這種從下而上的方法確保了平台具備日常工作所需的功能和特點，更貼近真實使用的情境。」

將BI Dataland這項高度複雜的戰略計畫的內部運作，作一個簡化的說明

端到端的數據生態系統？

• 資料來源：在確定要建立的內容之後，首先需要識別包含必要組件的方塊。
• 原始資料：其次，將所有方塊倒空，並將其中的組件收集到一個地方，形成一個「原始」的堆疊。
• 資料淨化：一些組件可能損壞或不符合其他組件。透過「淨化」堆疊，這些組件被移除。
• 資料庋用 ：為了加快建構過程，組件被「庋用」，也就是按照大小、顏色和用途進行排序。
• 預組的資料：透過「預組」那些在最終組裝中需要的第一批組件，可以進一步簡化並加快建構的過程。
• 資料模擬與分析：最後，預先建立的部件和磚塊被用來組裝和完成計劃中的建築。

「最大的挑戰並不是數據量有多龐大。」Urgeles 先生說道：「而是數據的結構如何讓公司所有人都能輕鬆使用。」

然而，一個數據生態系統無論組織得多好或全面，都無法獨立運作。「我們的員工需要具備解讀數據的正確知識，」百靈佳殷格翰的中央數據科學部門負責人 Brigitte Fuhr說道：「這就是為什麼進行正確的培訓非常重要。」

為了達到這個目標，百靈佳殷格翰成立了「數據科學學院」。「無論是資深數據科學家還是初學者，該學院將根據不同經驗的使用者提供合適的培訓。」Fuhr女士說道。

基於數據的客戶互動策略

Timmo Andersen，百靈佳殷格翰的人類藥物處方區域負責人（Head of Human Pharma Regions at Boehringer Ingelheim）。

作為第一個上線的Dataland應用程式，人類藥物處方區域的Next Best Action (NBA) AICER旨在支援銷售和市場營銷。西班牙已有超過250名銷售員正在使用該程式，以找到最佳的客戶互動和保留策略。

該程式從多個來源抓取數據，並使用數據科學功能來優化員工的通話策略。根據Timmo Andersen的說法，NBA已經提升了百靈佳殷格翰的業績。

「自應用程式上線以來，我們的客戶轉化率增加了近2％，銷售增長率提高了20％，這相當於1100萬歐元的現金流量，更重要的是，讓許多客戶獲得了優化後的客戶體驗。」Andersen先生補充說道：「基於數據的流程，讓我們從對事實的洞察而非單純的意見來改善我們與客戶間的互動。」

選擇臨床研究的試驗地點

Lorna Hart博士，臨床可行性全球負責人及Pegasus專案負責人（Global Lead of Clinical Feasibility and Head of the Pegasus project）。

Dataland具備一個最大的優勢，即是它能從整個公司中提取數據，找到新的關聯性。例如，Dataland應用程式「Pegasus Site Identification」可以通過分析所選國家的績效指標和模擬病患的招募情況，識別最佳的臨床試驗研究場所，以便更好地規劃試驗時間表。

臨床可行性全球負責人兼Pegasus專案負責人的Lorna Hart博士表示：「這個應用程式使我們更容易找到適合我們研究的國家與研究者，並顯著地加快整個流程。」

而當涉及臨床試驗時，Dataland還有其他可能的應用。一個研究的發展，以最佳化呈現和查看生成的數據非常重要。其中，一個名為「臨床開發座艙（Clinical Development Cockpit）」的應用程式就是由此而設計的。

加速個人化醫療的發展

George Okafo，百靈佳殷格翰全球醫療數據和分析部門負責人（Global Director of the Healthcare Data and Analytics Unit at Boehringer Ingelheim）。

在制藥業中，加快和改善產品開發的同時降低風險是最大的挑戰之一。擁有高品質的數據，並能夠對其進行評估對應對這一挑戰至關重要。

例如，生物銀行擁有大量的寶貴數據。生物銀行是一個將患者的實際訊息以匿名轉化為數據的訊息資料庫，其中包含有關組織樣本等材料特性和從電子健康記錄中獲得的其他數據。直到現在，百靈佳殷格翰無法以方便和整合的方式充分利用這些外部數據資源。而Dataland通過各種方式提供了這種能力，使公司能夠以更個人化的方式加快新藥開發的過程。

George Okafo是百靈佳殷格翰全球醫療數據和分析部門負責人（屬於全球計算生物學和數字科學部門）。他和他的團隊正在開發友善使用者的「Healthcare Data Analytics and Disease Translational Accelerator」平台，將生物銀行中的數據提供給百靈佳殷格翰的所有人使用，該平台將在今年下半年上線。

Okafo先生表示：「這些見解將幫助我們開發個人化的藥物，並確保它們比以前更快地到達患者手中。」

DIGITALIZATION

More data for more users means more value

Because data-based decisions are better decisions, Boehringer Ingelheim is establishing a data ecosystem called “dataland.” It will help all employees make data-driven decisions and improve the overall effectiveness of the organization. The uses so far are already demonstrating the system’s value.

Ferran Urgeles,

Program Manager of the dataland initiative at Boehringer Ingelheim

Whether it’s recommending products or moving unwanted emails to a spam folder, algorithms are part of everyday life — and not only online but also in many areas of the analog world. Consider, for example, the amount of digital data generated and recorded at every stage of the modern industrial process — whether in production, communication, customer contact, accounting or research. Algorithms analyze this data to yield new insights — detecting patterns and making connections that a human brain would never identify because of the sheer volume of information involved.

“Technology gives us a better understanding of which processes have an impact on the company’s performance, and how,” explains Ferran Urgeles, Program Manager of the dataland initiative at Boehringer Ingelheim. The goal: better decisions in every business area and every step of operations, from research to production, sales and beyond. “This will let us unleash the full potential of dataland that will provide insights we would have never discovered otherwise.”

And of course, the ultimate aim for Boehringer, Mr. Urgeles says, is “to help us develop medicines faster and more reliably.”

A secure, accessible data ecosystem

The dataland ecosystem is driven by a digital platform that collates data from every area of the company and makes it immediately available for uses like simulations and data analyses in a manner that is easy to comprehend.

“All of the data in the dataland ecosystem is put through data security and compliance checks and is protected using cutting-edge security measures,” Mr. Urgeles says. The ecosystem, which has been up and running at Boehringer Ingelheim since the end of 2022, revolves around practical use cases. This bottom-up approach ensures that the platform contains features and functions relevant on a day-to-day basis.

A simplification of the inner works of the highly complex strategic initiative BI dataland

The end-to-end data ecosystem?

• Data Sources : After knowing what to build, boxes which contain the necessary bricks have to be identified first.
• Raw Data : Second, all boxes are emptied, and the contained bricks collected in a single place as a „raw“ pile.
• Cleansed Data : Some bricks are broken or don’t fit to the others. By „cleansing“ the pile, these are removed.
• Curated Data : To speed up the building process the bricks are „curated“, this means to sort them by size, color and usage.
• Pre-assembled Data : Building can be further simplified and sped up by „pre-assembling“ first pieces that are needed for final assembly.
• Data Simulation & Analytics : Finally, the pre-built pieces and bricks are used to assemble and finalize the planned building.

“The biggest challenge is not so much the huge amount of data available,’’ Mr. Urgeles says. “It's about how it is structured so that it can be used easily by anyone in the company.”

But a data ecosystem cannot work on its own, no matter how well organized or comprehensive it is. “Our employees need to have the right skills to interpret the data,” says Brigitte Fuhr, Head of Boehringer Ingelheim’s Central Data Science Department. “That’s why proper training is so important.”

To achieve this end, Boehringer Ingelheim set up the “Data Science Academy.” “The academy provides training for users of any experience level – from veteran data scientists to people just getting started,” explains Mrs. Fuhr.

A data-driven approach to customer engagement

Timmo Andersen, Head of Human Pharma Regions at Boehringer Ingelheim

Next Best Action (NBA) AICER of Human Pharma Regions is the first dataland application to go live, developed to support sales and marketing. More than 250 sales employees in Spain are already using the program to find the best approach to customer engagement and retention.

The program draws in data from a variety of sources and uses data science functions to optimize the calling plans of the employees. NBA has already boosted Boehringer Ingelheim’s performance, according to Timmo Andersen, Head of Human Pharma Regions at Boehringer Ingelheim.

“Our customer conversion rate is up by almost two percent since the application went live,’’ Mr. Andersen says, “while the sales growth rate has gone up by 20 percent. That is equivalent to a net present value of 11 million euros, and even more importantly gives a lot of customers a next level of optimized customer engagement.”

Data-driven improvements to processes, Mr. Andersen adds, “allow us to collaborate on customer engagement based on insights — not on opinions.”

Selecting sites for clinical studies

Dr. Lorna Hart, Global Lead of Clinical Feasibility and Head of the Pegasus project

One of the biggest strengths of dataland is the way it pulls data from across the company to find new connections. Take the dataland application “Pegasus Site Identification.” It identifies the best possible sites for a clinical trial, by analyzing performance metrics of the countries under consideration and simulating patient enrollment for better planning of trial timelines.

“This application makes it a lot easier for us to identify the most suitable countries and investigators for our studies, and will speed up the process significantly,” explains Dr. Lorna Hart, Global Lead of Clinical Feasibility and Head of the Pegasus project.

And dataland has many more possible applications to offer when it comes to clinical trials. Once a study is underway, it’s important to find the best possible way to present and view the data that it generates. An application called the “Clinical Development Cockpit” is designed to do just that.

Speeding up personalized medicine development

George Okafo, Global Director of the Healthcare Data and Analytics Unit at Boehringer Ingelheim

Accelerating and improving product development, while reducing risks, is one of the biggest challenges for the pharmaceuticals sector. Having access to high-quality data, and the ability to evaluate it, is crucial to meeting this challenge.

Biobanks, for instance, contain huge amounts of valuable data. A biobank is a digital database of actual patient information — anonymized, of course — about the properties of materials such as tissue samples and other data gleaned from electronic health records. Until now, Boehringer Ingelheim has been unable to fully exploit these external data sources in a streamlined and integrated manner. Dataland provides this capability in ways that will enable the company to accelerate the process of developing new medicines in a more personalized manner.

George Okafo is the Global Director of the Healthcare Data and Analytics Unit (part of global Computational Biology and Digital Sciences) at Boehringer Ingelheim. He and his team are developing the user-friendly “Healthcare Data Analytics and Disease Translational Accelerator” platform to make the data held in biobanks available to everyone at Boehringer Ingelheim. It goes live later this year.

“These insights,” Mr. Okafo says, “will help us develop personalized medicines and make sure that they reach patients faster than before.