AUTHORS: DAGMAR HEBENSTREIT, DR. JÖRK HEBENSTREIT, DR. MICHAEL SCHOEMAKER
CAN THE TERMS MEDICAL PRODUCTS, REGULATED ENVIRONMENT, AND AGILE DEVELOPMENT METHODS BE UNIFIED, EVEN IN A UNIQUE TRADITIONAL GROUP ENVIRONMENT? THE INTUITIVE REACTION OF MANY OF THOSE ASKED IS PROBABLY A CLEAR NO, OR AT LEAST A DOUBTING FACIAL EXPRESSION. ROCHE DIABETES CARE TOOK UP THIS CHALLENGE AT THE END OF 2017 AND IMPLEMENTED AN INDIVIDUAL SCRUM SCALING MODEL.
At the end of 2017, Roche Diabetes Care started developing an innovative product that revolutionised the cash cow market segment of blood glucose monitoring. The question posed was: How can new, previously unprocessed market segments be opened up, and at the same time digitise blood sugar measurements?
The high degree of novelty of the market, as well as the technology, prompted Roche Diabetes Care to break new ground with the project setup, and to set up an agile R&D pilot project using SCRUM methodology. This was the only way conceivable to tackle the special challenges, despite the established traditional development processes and organisational structures. Short iteration cycles and quick decisions on how to proceed quickly lead to new insights and enabled the first product launch within a short project term.
The question posed was: How can new, previously unprocessed market segments be opened up, and at the same time digitise blood sugar measurements?
Roche received support from a network of external coaches. This first test balloon helped the company gain experience with agile organisational approaches and to learn more about its own hurdles and needs. Roche Diabetes Care had to struggle, amongst other things, with long decision-making processes, organisational hurdles and an over-regulating process landscape. After the launch of a Minimal Viable Product (MVP) at the end of 2018, the course was set, and Roche Diabetes Care decided to start further development projects using an agile setup.
As a result, a decision was made to use a Scaled Agile Framework (SAFe) on which adjustments were made.
When Roche Diabetes Care decided to start larger product developments with more than one development team, there was a need for a suitable framework for the required scaling. This was intended to better manage overall control, prioritisation and resource allocation in the overall development. Together with the already established consultants, various known scaling frameworks were intensively analysed in several workshops and evaluated using Roche-specific requirements.
An innovative micro insulin pump, launched in pilot markets in mid-2017, has to date, been further developed by a large number of independent project teams in all its aspects, using the traditional linear waterfall model. Different budgets, project managers and management oversight bodies made the joint project work noticeably more difficult. There was a lack of transparency and overriding priorities. In addition, employees were assigned to several projects at the same time, which led to efficiency losses due to multi-tasking and lack of focus. At the same time, it related to a complex class 2a medical device with low fault-tolerant mechanics, electronics and embedded firmware in a highly competitive market environment.
How do you convert such a complex project that has already started, with several independent project teams in the ongoing development process, to a completely new method?
In order to create a common know-how basis, Roche Diabetes Care invested in intensive training for future teams (approx. 50 employees) within the programme over a period of 3.5 months. Basic in-house training was carried out to understand agile working methods, training within the SAFe scaling model, as well as special training for the agile development of hardware components and the fulfilment of regulatory requirements for iterative development (according to AAMI TIR 45). In addition, potential candidates for Scrum Master, Product Owner and Architect roles, completed special training and further education. During the entire project setup, management was intensively involved in decisions on the overall organisation, new roles, appointments and processes.
Prior to the official go-live of the agile organisation, work was carried out on prerequisites, such as setting sprint and increment lengths, Definitions of Done within the teams, in accordance with the requirements and deliverables of the internal quality management system.
The challenges of the change-over are particularly evident in cooperation with other business areas, which are still very traditional with the linear development model. The internal quality management system in particular prescribes many linear development steps, which make incremental development significantly more difficult. Many medical technology companies have over-regulated themselves by implementing linear models that are not required by the authorities in this form. The approval authorities expect complete documentation on medical devices but give no opinion on the development method to be used, i.e. the way in which this documentation is created.
As with any agile transformation, it takes some time to achieve the transfer of responsibility from management to the teams, and in particular to the role of product owner. New product owners are also faced with the task of maintaining complex backlog structures and prioritising them in permanent coordination with stakeholders.
In particular, a lack of skills with respect to this new role and insufficient time allocated to it can be a problem at first. Accompanying coaching of the role is necessary during this personal development. Sole assignment of employees to a development team has not yet proven successful in all cases. Furthermore, the first agile programmes are always faced with challenges that arise from working with the still existing traditional corporate environment. Therefore, surrounding areas must be intensively integrated and incorporated, even if they have not yet been integrated into the agile way of working.
Nevertheless, the teams found their way around very quickly and gained a lot of speed as a result of focus and clear priorities. The high level of transparency within the organisation is viewed by the stakeholders in particular as high added value. In addition, stakeholders, as well as project members, appreciate the short and regular feedback cycles and personal responsibility of the development teams.
Despite all the challenges that still exist, the results, measured in terms of implementation time, are outstanding and extremely satisfactory. The development of team dynamics, team spirit and motivation within the programme are particularly impressive. Increased employee motivation associated with the change and consolidation of the way of working can be directly measured by higher team productivity. In particular, the experience of self-organisation, increased freedom of choice, assumption of personal responsibility, transparency with respect to successes, obstacles and progress, as well as the direct connection of employees own work to the “big picture” lead to a significantly higher level of satisfaction among many.
The development of team dynamics, team spirit and motivation within the programme are particularly impressive.
Don’t do this at home! In an initial approach, before the end of 2017, Roche Diabetes Care had already taken steps on its own towards an agile way of working in software development and ended up with a SCRUM waterfall hybrid model, which combined the disadvantages of both worlds and generated little benefit and acceptance within the organisation. Competent support with respect to the setup and consistent support during implementation by experienced coaches are a key success factor. It was particularly important to Roche Diabetes Care to have know-how for all required business areas on board and to start with a network of experienced coaches who were all experts in their field . This resulted in a unique combination of hardware, medical device regulation, software and change management know-how. This network takes a 360° perspective and everyone within it empowers each other by building on close content and methodological coordination.
Such a big change can only be achieved within a short time if all participants are involved and at the same time become multiplicators of the new setup and mindset.
Intensive coaching of the teams is essential during the implementation phase. Such a big change can only be achieved within a short time if all participants are involved and at the same time become multiplicators of the new setup and mindset. At the same time, the organisation is taken over by a separate transition team that accompanies the change process. The most important thing here is the corresponding management support. A lack of support by managers, together with a lack of methodological knowledge and disregard for agile values, are the top reasons for agile failures.
Agile development methods in a highly regulated environment are possible, including in the area of hardware development. Awareness of the challenges is important in order to identify some of the stumbling blocks in advance. Real agile transition always includes cultural change within an organisation that goes far beyond the purely technical introduction of a methodology. Every change in behaviour takes time and requires the opportunity to be tried, accepted and approved. You should allow employees this time.
Despite all the successes, there are also obstacles, such as the fact that agile and traditionally managed projects are running parallel to each other, or the areas outside of development and product management are still working in their established mode and there is a lack of understanding of the change and the fact that the processes could not yet be fully adapted to the agile way of working. This is where the company has some catching up to do.
Experience from this ongoing project shows, however, that with people who are willing to engage in new things, try new methods, critically question things, organise themselves and take real responsibility, successful transitions are actually possible. Roche Diabetes Care is on an extremely good and promising path.
 Network of consultants from Comeno GmbH, parsQube GmbH, DENKRICHTUNGEN, Agile Quality Systems LLC and AGILEUS Consulting GmbH & Co. KG.
Unbedingt erforderliche Cookies ermöglichen grundlegende Funktionen und sind für die einwandfreie Funktion der Website erforderlich. Daher kann man sie nicht deaktivieren. Diese Art von Cookies wird ausschließlich von dem Betreiber der Website verwendet (First-Party-Cookie) und sämtliche Informationen, die in den Cookies gespeichert sind, werden nur an diese Website gesendet.
Cookielaw Dieses Cookie zeigt das Cookie-Banner an und speichert die Cookie-Einstellungen des Besuchers. Anbieter: Jimdo GmbH, Stresemannstrasse 375, 22761 Hamburg, Deutschland. Cookie-Name: ckies_cookielaw Cookie-Laufzeit: 1 Jahr Cookie-Richtlinie: https://www.jimdo.com/de/info/cookies/policy/ Datenschutzerklärung: https://www.jimdo.com/de/info/datenschutzerklaerung/
Jimdo Control Cookies Steuerungs-Cookies zur Aktivierung der vom Website-Besucher ausgewählten Dienste/Cookies und zur Speicherung der entsprechenden Cookie-Einstellungen. Anbieter: Jimdo GmbH, Stresemannstraße 375, 22761 Hamburg, Deutschland. Cookie-Namen: ckies_*, ckies_postfinance, ckies_stripe, ckies_powr, ckies_google, ckies_cookielaw, ckies_ga, ckies_jimdo_analytics, ckies_fb_analytics, ckies_fr Cookie-Laufzeit: 1 Jahr Cookie-Richtlinie: https://www.jimdo.com/de/info/cookies/policy/ Datenschutzerklärung: https://www.jimdo.com/de/info/datenschutzerklaerung/
Funktionelle Cookies ermöglichen dieser Website, bestimmte Funktionen zur Verfügung zu stellen und Informationen zu speichern, die vom Nutzer eingegeben wurden – beispielsweise bereits registrierte Namen oder die Sprachauswahl. Damit werden verbesserte und personalisierte Funktionen gewährleistet.
POWr.io Cookies Diese Cookies registrieren anonyme, statistische Daten über das Verhalten des Besuchers dieser Website und sind verantwortlich für die Gewährleistung der Funktionalität bestimmter Widgets, die auf dieser Website eingesetzt werden. Sie werden ausschließlich für interne Analysen durch den Webseitenbetreiber verwendet z. B. für den Besucherzähler. Anbieter: Powr.io, POWr HQ, 340 Pine Street, San Francisco, California 94104, USA. Cookie Namen und Laufzeiten: yahoy_unique_[unique id] (Laufzeit: Sitzung), POWR_PRODUCTION (Laufzeit: Sitzung), ahoy_visitor (Laufzeit: 2 Jahre), ahoy_visit (Laufzeit: 1 Tag), src (Laufzeit: 30 Tage) Security, _gid Persistent (Laufzeit: 1 Tag). Cookie-Richtlinie: https://www.powr.io/privacy Datenschutzerklärung: https://www.powr.io/privacy
Die Performance-Cookies sammeln Informationen darüber, wie diese Website genutzt wird. Der Betreiber der Website nutzt diese Cookies um die Attraktivität, den Inhalt und die Funktionalität der Website zu verbessern.
Jimdo Statistics Die sog. Jimdo Statistikfunktion ist eine Tracking-Technologie, basierend auf Google Analytics, welche von der Jimdo GmbH betrieben wird. Anbieter: Jimdo GmbH, Stresemannstrasse 375, 22761 Hamburg, Deutschland. Cookie-Namen und Laufzeiten: ckies_jimdo_analytics (Laufzeit: 2 Jahre), __utma (Laufzeit: 2 Jahre), __utmb (Laufzeit: 30 Minuten), __utmc (Laufzeit: Sitzung), __utmz (Laufzeit: 6 Monate), __utmt_b (Laufzeit: 1 Tag), __utm[unique ID] (Laufzeit: 2 Jahre), __ga (Laufzeit: 2 Jahre), __gat (Laufzeit: 1 Min), __gid (Laufzeit: 24 Stunden), __ga_disable_* (Laufzeit: 100 Jahre). Cookie-Richtlinie: https://www.jimdo.com/de/info/cookies/policy/ Datenschutzerklärung: https://www.jimdo.com/de/info/datenschutzerklaerung/
Google Analytics Diese Cookies sammeln zu Analysezwecken anonymisierte Informationen darüber, wie Nutzer diese Website verwenden. Anbieter: Google LLC, 1600 Amphitheatre Parkway, Mountain View, CA 94043, USA oder Google Ireland Limited, Gordon House, Barrow Street, Dublin 4, Irland, wenn Sie in der EU ansässig sind. Cookie-Namen und Laufzeiten: __utma (Laufzeit: 2 Jahre), __utmb (Laufzeit: 30 Minuten), __utmc (Laufzeit: Sitzung), __utmz (Laufzeit: 6 Monate), __utmt_b (Laufzeit: 1 Tag), __utm[unique ID] (Laufzeit: 2 Jahre), __ga (Laufzeit: 2 Jahre), __gat (Laufzeit: 1 Min), __gid (Laufzeit: 24 Stunden), __ga_disable_* (Laufzeit: 100 Jahre). Cookie-Richtlinie: https://policies.google.com/technologies/cookies Datenschutzerklärung: https://policies.google.com/privacy
Marketing- / Third Party-Cookies stammen unter anderem von externen Werbeunternehmen und werden verwendet, um Informationen über die vom Nutzer besuchten Websites zu sammeln, um z. B. zielgruppenorientierte Werbung für den Benutzer zu erstellen.
Cookies werden zur Benutzerführung und Webanalyse verwendet und helfen dabei, diese Webseite zu verbessern. Durch die weitere Nutzung dieser Webseite erklären Sie sich mit unserer Cookie-Richtlinie einverstanden.