Convergent Innovation in Emerging Healthcare Technology Ecosystems: Addressing Complexity and Integration

Precision Medicine and Digital Health are increasingly important areas that are reliant on “convergent” or “cross-industry” innovation (Sabatier et al. 2012; Thakur et al., 2012). A consequence of convergence is that it brings more uncertainty and allows greater influence from new knowledge and actors, including previously disparate technologies and capabilities (Rikkiev & Mäkinen, 2013). In turn, there is an added complexity because convergence contradicts the two dominant forms of organizational learning, namely simplification and specialization (Levinthal & March, 1993). This research focuses on the uncertainty and complex integration issues that arise from the emerging ecosystem, from developing the innovation and in forming a viable value network. Much of the extant innovation literature has focused on innovation by incumbents in existing industries or with existing value-chain partners (Enkel & Gassmann, 2010). More recently, there has been increasing interest in “cross-industry” or “convergent” innovation (Gassmann et al. 2010; Stieglitz, 2003). However, convergence can result in higher levels of equivocality, uncertainty, and risk as the diverse technology, alliance partners, and ecosystems merge (Enkel & Heil, 2014; Hacklin, 2005; Mason et al., 2013). These considerations manifest themselves as different integration challenges that depend on the nature of the convergence (Rikkiev & Mäkinen, 2013).


Introduction
Precision Medicine and Digital Health are increasingly important areas that are reliant on "convergent" or "cross-industry" innovation (Sabatier et al. 2012;Thakur et al., 2012).A consequence of convergence is that it brings more uncertainty and allows greater influence from new knowledge and actors, including previously disparate technologies and capabilities (Rikkiev & Mäkinen, 2013).In turn, there is an added complexity because convergence contradicts the two dominant forms of organizational learning, namely simplification and specialization (Levinthal & March, 1993).This research focuses on the uncertainty and complex integration issues that arise from the emerging ecosystem, from developing the innovation and in forming a viable value network.
Much of the extant innovation literature has focused on innovation by incumbents in existing industries or with existing value-chain partners (Enkel & Gassmann, 2010).More recently, there has been increasing interest in "cross-industry" or "convergent" innovation (Gassmann et al. 2010;Stieglitz, 2003).However, convergence can result in higher levels of equivocality, uncertainty, and risk as the diverse technology, alliance partners, and ecosystems merge (Enkel & Heil, 2014;Hacklin, 2005;Mason et al., 2013).These considerations manifest themselves as different integration challenges that depend on the nature of the convergence (Rikkiev & Mäkinen, 2013).
For convergence in healthcare technologies, apart from several practitioner articles (Eselius et al., 2008;Gupta et al., 2013;Mason et al., 2013), there are limited studies Precision Medicine and Digital Health are emerging areas in healthcare, and they are underpinned by convergent or cross-industry innovation.However, convergence results in greater uncertainty and complexity in terms of technologies, value networks, and organization.There has been limited empirical research on emerging and convergent ecosystems, especially in addressing the issue of integration.This research identifies how organizations innovate in emerging and convergent ecosystems, specifically, how they address the challenge of integration.We base our research on empirical analyses using a series of longitudinal case studies employing a combination of case interviews, field observations, and documents.Our findings identify a need to embrace the complexity by adopting a variety of approaches that balance "credibility-seeking" and "advantage-seeking" behaviours, to navigate, negotiate, and nurture both the innovation and ecosystem, in addition to a combination of "analysis" and "synthesis" actions to manage aspects of integration.We contribute to the convergent innovation agenda and provide practical approaches for innovators in this domain.

" "
Between the idea And the reality Between the motion And the act Falls the Shadow In The Hollow Men (1925) that examine the implications for technological or business model discontinuities (Bojovic et al., 2015;Sabatier et al., 2012) and these few (Bernabo et al., 2009;Dubé et al., 2014;Ramachandran et al., 2011;Shmulewitz et al., 2006) focus more on the phenomenon than on the implications.
Using empirical analyses in five longitudinal case studies with a combination of interviews, field observations (e.g., meetings and workshops), and documents, our exploratory research findings point to a need to embrace the complexity.We propose the adoption of approaches that balance taking "credibility-seeking" and "advantage-seeking" positions using non-ergodic routines that navigate, negotiate, and nurture with a combination of "analysis" and "synthesis" actions to manage integration.

Theoretical Background
Addressing uncertainty and complexity Uncertainty and risk are inherent in innovation and arise from four types of complexity: evolutionary, temporal, relational, and cultural (Garud et al., 2013).Importantly, there are inherent differences between managing risks (with known probabilities) and uncertainty (or "unknown unknowns") (Teece et al., 2016).
The major uncertainties and risks in innovation are generally considered to be technological, regulatory, and market based (Hobday, 1998), and they are typically addressed by a variety of mechanisms to "manage complexity", resulting in simplification and specialization (Levinthal & March, 1993).However, such approaches create limitations and may inhibit the innovation itself (Garud et al., 2013).Although several of these challenges are acknowledged (Rikkiev & Mäkinen, 2013), there has been limited empirical research to understand how they are addressed.
Differences between the nature of innovation and its impact have been considered in both the innovation literature (Abernathy & Clark, 1985) and the diffusion literature (Rogers, 2003).In extant literature, there is more focus on the management of risk (Evanschitzky et al., 2012) than on addressing uncertainty, which is considered more likely and harder to manage (Teece et al., 2016).
In addressing uncertainty, McGrath (2001) confirms the earlier findings of March (1991) that the degree of exploration is important; broader searches across more variety can improve performance.The dynamic capab-ility literature points to the use of sensing, seizing, and transforming to better manage uncertainty (Teece et al., 2016), with abduction (as a mode of inference) being important to create new thinking for subsequent testing.This suggests creative abduction (Schurz, 2008) is more relevant (versus selective abduction, which chooses from multiple explanations), although creative abduction is rarely discussed in the literature (Prendinger & Ishizuka, 2005).
Sommer and colleagues ( 2009) identify two approaches to respond to uncertainty: selectionism and trial-and-error.Selectionism refers to attempting many solutions in parallel and selecting the best based on the outcomes.However, such an approach can be costly and potentially inefficient.Trial-and-error learning refers to adjusting activities and targets as new information becomes available.The combination of complexity and uncertainty, and the need for creative and exploratory approaches using limited and often equivocal information, is counter to much of the traditional innovation literature with linear processes and defined decision criteria, as highlighted by Garud and colleagues (2013) and Bessant and colleagues (2005).

Integration challenges
Integration, by (re)combining knowledge, is inherent in innovation (Grant, 1996;Kogut & Zander, 1993;Teece, 1996).As well as knowledge or technology integration, there is a need for market and organization integration (Tidd & Bessant, 2013).Much of the "integration" literature focuses on intra-organization and cross-functional integration as Evanschitzky and colleagues (2012) identified in their meta-analysis of success factors in 233 innovation studies.Although integration (internal and external) has been identified as an indicator of innovation performance, it is moderated by equivocality (Koufteros et al. 2005).Yet, equivocality is itself inherent in convergence.
Alliance formation (Colombo et al., 2006;Eisenhardt & Schoonhoven, 1996) and management under conditions of high uncertainty would, therefore, appear to be a critical capability for startups and new ventures within an incumbent firm.Previous literature has identified the need for a highly integrated value network as a key factor in performance (Prajogo & Olhager, 2012), but this presupposes a strong understanding of the needs and capabilities of the alliance partners.In convergent innovation, ecosystems and value networks are emerging, so a comprehensive understanding may be lacking.

Convergent Innovation in Emerging Healthcare Technology Ecosystems
Mark A. Phillips, Tomás S. Harrington, and Jagjit Singh Srai Systems-integration risks are not new (for example, see Henderson & Clark, 1990), but have traditionally been addressed by concepts such as modularity (Baldwin & Clark, 1997;Schilling, 2000).However, the presumption in such an approach is that the knowledge is well codified (Cardinal et al., 2001).In convergence, this is more challenging, because such codification is initially limited.
What is less clear in the extant literature is how this complexity and integration is addressed.Garud and colleagues ( 2013) identified some challenges and resulting gaps in both research and practice, and they call for approaches that embrace the complexity as a "generative" process, rather than trying to simplify and "manage" it.

Research Design
This research aims to address these issues by considering the question of how organizations address the challenges of integration in convergent technology innovation within the wider context of convergent innovation for healthcare and medical technologies in emerging ecosystems.
Given the context of the enquiry, and the evolving nature of the setting, a qualitative approach was adopted (Yin, 2014).The design consisted of two main phases (see Figure 1).An exploratory phase involved 27 semi-structured interviews from a wide range of ecosystem stakeholders, which enabled better understanding of the emerging ecosystem itself (Table 1).The interviews were analyzed inductively using the Gioia (2012) method to identify "dimensions".From these dimensions and a review of innovation and ecosystem literature, an investigational tool was developed (using abduction) for use in the second phase.The second

Convergent Innovation in Emerging Healthcare Technology Ecosystems
Mark A. Phillips, Tomás S. Harrington, and Jagjit Singh Srai

Convergent Innovation in Emerging Healthcare Technology Ecosystems
Mark A. Phillips, Tomás S. Harrington, and Jagjit Singh Srai phase was based on empirical analyses of five in-depth longitudinal case studies conducted over 15-to 24month periods employing a combination of interviews, field observations, and primary documents (obtained under confidentiality) as data sources, together with supplementary evidence from public documents.The cases involved three established companies and two startups, with 62 case study interviews, 41 observations, and over 100 documents (see Table 2).Further ecosystem interviews were also conducted to provide contemporaneous context.The data were collected and analyzed using thematic and process coding to identify patterns.A further in-depth analysis based on Sayer's (1992) approach was then used to identify the potential underlying causal mechanisms using the ecosystem data as context (conditions and constraints).

Findings
The exploratory ecosystem interviews identified major issues for actors in understanding the ecosystem itself, the diverse perspectives of actors, and how to create and capture value.But the ecosystem not only creates "problems", it also provides "solutions" for innovators.
There is therefore an explicit link between the ecosystem, the innovation, and capabilities needed.
All the cases provided evidence that organizations undertook activities to search and sense-make (and sensegive) in the emerging ecosystem.But the nature of those search and sense-making activities differed; those adopting a more exploratory and engaging approach, for example by snowballing (Goodman, 1961) to identify

Convergent Innovation in Emerging Healthcare Technology Ecosystems
Mark A. Phillips, Tomás S. Harrington, and Jagjit Singh Srai distant actors and then engaging them, appear to be more successful.The case findings point to extensive, repeated, and direct interactions as important for sense-making.Decision-making processes were largely informal (and invariably supported by external expertise), using directional criteria, and focused on key issues in terms of balancing value and risk.
Given the newness of the ecosystem, firms invested in activities that aided understanding and created credibility among potential partners, enabling them to engage, negotiate, and move to a position of advantage-seeking.However, these efforts were balanced by activities that continued to support or sustain the ecosystem itself, often with no immediate return, as described by the leader of one case (DH1): "…there needs to be 'congruence', a real alignment.Not just in terms of the outcome, but also cultural and how you are going to do it.Connections do not just happen -you need to 'cultivate' to create the right opportunities." The uncertainty in the ecosystem presents issues, but is also a potential source of solutions.The casual mechanism analysis, derived from Sayer (1992), suggests organizations need to "navigate" the ecosystem, "negotiate" a position, and "nurture" the innovation by a combination of "credibility-seeking" and "advantage-seeking" activities that are "generative" in that they create opportunities.These activities appear to be underpinned by five interrelated processes or organizational routines: searching, sense-making, selecting, shaping, and sustaining.A series of findings and insights from our case studies are summarized in Table 3.These activities and routines support four main objectives to shape the innovation and create value, to manage risks and the integration, and to develop the value network and wider ecosystem (Figure 2).

Discussion and Implications for Practice
The integration problem, as identified earlier, is complex and does not just include technological or market risk, but requires a simultaneous balancing of risk around four aspects: i) technical systems integration, ii) commercial or business models, iii) value network, and iv) organizational integration (O'Connor & Rice, 2013).

Technical systems and integration risks
The

Mark A. Phillips, Tomás S. Harrington, and Jagjit Singh Srai
To overcome these challenges, all the cases worked in a collaborative way with other knowledge and alliance partners, creating opportunities to understand and share.This finding suggests that the approach appears more dependent on building relationships, rather than on information codification (Tidd & Bessant, 2013) and traditional technology integration approaches.

Market and business model risk
Convergent innovation, with increasingly digital content, provides opportunities for innovators to disrupt existing health and care pathways, making the identification of the value proposition and customer more complex and riskier.The nature of the technology used by Case MLD provided multiple options for business models, providing a "platform" from multiple revenue streams.Similarly, Case NMD identified several business models that might be appropriate depending on the success of the technology and its clinical application.However, such changes are not evident from the outset and do not appear to be readily designed, as they often emerge and evolve along with the innovation.

Value network risk
The prevalent approach from the cases was to first build transient partnerships.In doing so, the case firms developed knowledge and built relationships over time, thereby reducing risks.More robust relationships and long-term alliances were developed later.There is a 'trading off' of some short-term risk (by not having wellestablished networks) against making a "bad decision" on a longer-term partner.The alternative -to delay the formation of any partnerships and thus delay the innovation itself -was also observed in Case DH2, which ultimately was a failed venture.

Internal organizational risks
The risk of an innovation not being accepted by the incumbent organization is widely accepted in the literature (e.g., Danneels, 2011).To avoid resistance and mitigate organizational risk, the cases made multiple but small changes to existing routines.Examples of this approach were identified in Cases NMD, CMTI, and DH1.

Summarizing approaches for addressing complexity and integration
Risks arose from multiple sources: these risks could be considered in isolation, but they are interrelated.They form elements of a complex system, but rather than attempting to simplify the system, it is suggested that the complexity is more often addressed in a holistic way.
For example, Case MLD undertook multiple risk reviews, whereby, they address patient and user risks, technology risk, business model risks, and overall project management risks.Similarly, Case NMD took a systemic approach to managing risks, and having mapped the major risk areas at an early stage, they set about addressing those risks in multiple areas (including for example understanding the human biology, developing human-machine interfaces, developing new energy systems, and developing new ways to interpret novel data).
The evidence suggests a move beyond the multiple risk approaches identified for disruptive innovation (e.g., Keizer & Halman, 2007) to more comprehensive models as proposed by O'Connor and Rice (2013).
Despite knowing these represent categories of uncertainty that need to be addressed, it does not answer the core question -how?Revisiting the case evidence indicates several approaches being employed.Some are rooted in process, for example, in conducting formal risk assessments (as in Cases MLD and NMD) and making changes to processes to minimize or mitigate risk (as in cases NMD and DH1).Others aimed at building relations (evident in the Cases NMD, MLD, DH1, and CMTI, as previously discussed).Finally, there are cases that are more elusive and harder to classify, but are broadly based around management decisions and propensity to address wider ecosystem risks or in shaping the innovation "agency".
An early treatise on innovation by Usher (1966), revised from a book originally written in 1926, identifies two types of action that innovators may use: analytic (analysis) and synthesis.Analytic approaches can be conceived as using systematic methods to address largely anticipated or perceived gaps.Synthesis approaches are more creative and look to position the innovation to take advantage of future options.Revisiting the activity system suggested earlier (Figure 2), the underpinning routines may be conceived as being either largely analysis or largely synthesis driven by either a process, relational, or agency focus.This view suggests a conceptual model (Figure 3) that aims to position the integrating and risk management activities, in context, with the underlying approach.
This view also suggests approaches that are nondeterministic.Equally, they are not arbitrary, but nonergodic (Sydow et al., 2012).In such a complex ecosystem, it is unlikely that any previous state will be re-experienced and, hence, innovation approaches become more context sensitive.The challenge is in embracing

Convergent Innovation in Emerging Healthcare Technology Ecosystems
Mark A. Phillips, Tomás S. Harrington, and Jagjit Singh Srai the complexity and managing the integration.Importantly, the innovator should not fixate too much on any one of the axes in Figure 3, but should look to "flex" between analytic and synthesis actions as needed and as opportunities arise.
There appears from the cases to be no single way of organizing, but they suggest a combination of activities and capabilities to access information and partners, to respond to technological, organizational, and ecosystem changes, and to maintain a focus on outcomes and performance.

Conclusion
The case evidence suggests innovators should undertake multiple engagements with diverse stakeholders as part of a search, sense-making, and selection process.
Critically, this process can also help to create credibility and visibility within the ecosystem -necessary precursors to form alliances and create opportunities to achieve first-mover advantage.Innovators also have an opportunity to shape outcomes and their value network, but the importance of supporting and sustaining the emerging ecosystem is also identified here as a key activity.
Activities to sustain and support an innovation (or to shape it) are largely a result of management agency -to identify opportunities or challenges and then act to address them.The development of credibility, and later advantage-seeking positions, are the result of relational activities.The physical creation of value, integration, and the reduction of risk are primarily process driven.Actions to sustain, to seek credibility, and to reduce risk are effected by analytic approaches (analysis), in assessing the current state, developing options, and then deciding the best course.Finally, the value creation, advantage-seeking, and shaping activities are more about synthesis -identifying opportunities in patterns as they emerge.
This exploratory research addresses a relatively new phenomenon and so is limited to a few cases, therefore, limiting the generalizability.A qualitative approach was used, but despite significant observations and interviews, risk remains in inference and interviewee reliability.Our cases are focused on the United Kingdom but also involve partners from outside the UK.Although the cases are longitudinal, they were only studied for two years; however, they represent a formative part of the specific innovations and include major decisions or pivot points.
Future research would point to the need to better understand the emergence of such ecosystems and their impact on innovator processes in different contexts (e.g., different convergence regimes).
In summary, convergent innovation brings increased complexity and integration challenges that are not deterministic.There is a need to "embrace the complexity" by adopting a variety of approaches that balance credibility-seeking and advantage-seeking behaviours

Convergent Innovation in Emerging Healthcare Technology Ecosystems
Mark A. Phillips, Tomás S. Harrington, and Jagjit Singh Srai and oscillate between analysis and synthesis actions to address technological system, market, organizational, and value network integration risks.Although limited to a few cases in an emerging ecosystem, by taking a contemporaneous and longitudinal case approach, we address an identified gap in the literature on "how" organizations innovate in this context.
bringing together of different scientific, technical, and industry knowledge inevitably adds a new dimension to the technical risk -that of technical systems integration.The cases highlighted several examples: Case NMD sought to integrate diverse science and technology from biology, micro-electronics, flexible electronics, new neural interfaces, energy harvesting (all at a much smaller scale than previously conceived), and new control algorithms.Similarly, Case MLD integrated visual cognition science with "millisecond scale" response monitoring on mobile technologies, cloud computing, and artificial intelligence (AI) technology.

Figure 2 .
Figure 2. Proposed activity system model of convergent innovation

Figure 3 .
Figure 3. Integration of innovation activities

Table 1 .
Ecosystem interviews to develop context and constraints Figure 1.Overall research approach www.timreview.ca

Table 2 .
Case research sources www.timreview.ca Prior to joining the Faculty of Social Sciences at UEA in August 2017, Tomás spent eight years at the University of Cambridge's Institute for Manufacturing.His research and practice interests focus on industrial systems transformation, enabled by the adoption of advanced manufacturing and digital technologies.He has also held senior roles in industry encompassing new product development, process design, and big data analytics -most recently with Intel Corporation.Tomás holds Bachelor and PhD degrees in Chemistry and an MBA (with distinction) for which he received a Chartered Management Institute award in 2008.Jagjit Singh Srai is Head of the Centre for International Manufacturing within the Institute for Manufacturing at the University of Cambridge, United Kingdom.His research focuses on the analysis, design, and operation of international production, supply and service networks, and the disruptive impacts of new technologies, markets, and regulations.As Research Director of Project Remedies, a £23m collaborative research programme involving leading pharmaceutical firms, applied research explores how new technologies may transform healthcare supply chains.Jag also advises leading multinationals, governments, and international institutions including UNCTAD, UNIDO, and WEF.Previous roles have been in industry with Unilever working as a Supply Chain Director of a multinational regional business, Technical Director of a national business, and other senior management positions.He holds a first-class honours degree in Chemical Engineering from Aston University, United Kingdom, and MPhil and PhD degrees in International Supply Networks from Cambridge University, and he is a Chartered Engineer and a Fellow of the Institute of Chemical Engineers.