Translational Research Institute for Space Health (TRISH) FIP Proposal
Title: Sourcing Innovation from Startup Companies: Collaborative Health Innovation Platform
Investigators: Yael V. Hochberg, Ph.D., M.A., B.Sc.
Daniel Lee, Ph.D., M.A, B.A.
Performance Site: Rice University
Implementation Partner: Energizing Health
In this project we aim to develop a software-based platform (The Collaborative Health Innovation Platform, or CHIP henceforth). This platform will provide census-level information on healthcare startups and partner organizations. In addition it will serve as a tool for running randomized evaluations related to innovation sourcing. The platform will service both the research and practitioner communities.
Over the last three decades, many large corporations and government organizations have turned to an Open Innovation model to source new ideas and technologies to meet particular needs or to combine with existing knowledge to further new generation of products and services. While successful open innovation can lead to shared rewards, this new paradigm is not without its costs, as matching between startups and large organizations can entail significant difficulties. As such, there is a need to identify willing startups and facilitate the matchmaking process.
This need is particularly strong in the healthcare arena. Major research institutions and grant-making organizations, including, NASA, DARPA, the Translational Research Institute (TRISH) for Space Health, investors, nonprofit foundations, and hospital systems, must reach a global network of entrepreneurs and researchers to find, fund and deliver solutions. While relevant space-health startups may have the most inventive products, they may also be particularly capital constrained, and may be unaware of potential applications for their technologies or interest on the part of larger organizations. Similarly, even if large organizations can identify startups with relevant technologies, convincing those startups to agree to work with larger, more conservative and bureaucratic organizations can be difficult.
The need to identify startups in turn presents several interesting research questions. What kind of startups are likely to be willing to partner with large organizations and what characterizes them? What do these startups look for in partners? What obstacles do they face?
In this proposal, we propose the development of a Collaborative Health Innovation Platform, or CHIP—a software-based platform on which we propose to run hypothesis-based experiments to answer the above questions, and which, eventually, can be used to help to solve these search-and-sorting problems. We intend to use the CHIP platform to run hypothesis-based experimentation around startup partnerships with the ultimate aim of reducing search costs and improve matching between partner organizations. Further, by pairing CHIP with survey data, we aim to provide qualitative evidence on the concerns startups have in working with large organizations (and vice-versa) that can further inform and enable solution designs.
To a large extent, we don’t know what the space of innovation-driven healthcare startups looks like. Identifying startup companies with relevant technologies is difficult, due to the lack of a centralized database and sorting and identification tools. This lack of a coherent landscape also leaves open many questions about the choice of such startups to partner with large organizations. While this is in essence a 2-sided matching market, a robust and interactive platform for connecting ‘problem owners’ with ‘problem solvers’ does not exist.
In developing the CHIP platform, we aim to better understand this market by running randomized evaluations designed to alleviate market frictions, thereby reducing both monetary and bandwidth costs to both startups and the TRISH. Eventually, the platform will be opened to allow for corporations and other large organizations to use the knowledge we have gained from this project to more easily identify and partner with startups.
The question of how startups match to large organizations, when they are willing to work together, and what facilitates successful collaboration, remains a fundamental open question in the entrepreneurship and innovation literature. Many elements are thought to influence the decision to partner, including founder background, company traction and team capabilities. In particular, the question of what kind of startups are likely to be willing to partner with large organizations has not been addressed with causal rigor, while much of the existing theory about matching and strategic alliances (See e.g., Wang and Rajagopalan, 2015 for a review) is based on the choice of incumbent firms.
There is very little work on the choices startups face. This is in part because of the paucity of high-quality data, but also because much of the theory is based in industry verticals such as biopharma, where the choice to partner is not a question of “if” but rather one of “when”.
We address this shortcoming in several ways. First, through the development of a CHIP software platform. We envision CHIP as a unique one-stop-shop of collaboration, investment, and incubation for biomedical technologies. The platform is in effect a database populated from leading innovation hubs and companies and technologies from the existing professional networks of TRISH Development Specialists, as well as by our affiliated partner organizations. The database will be fully tagged and searchable by risks, gaps, innovators, academic groups, development stage, and more. The value added to the TRISH in seeding this platform is the continual stream of quality opportunities, and as the database expands, TRISH will improve its signal-to-noise ratio by sourcing higher quality startups and innovation.
Next, once this platform is developed, we will use it to run randomized evaluations to answer questions related to startup identification and motivation. That is: who is willing to engage, what they need to be able to engage, and knowing that, how to get them to engage.
A promising line of inquiry within these broad questions involves messaging campaigns. For instance, we can envision several reasons for responding when TRISH puts out a Request for Proposals (RFP). By randomizing various elements emphasized within an RFP announcement, we can begin to assess startup motivations behind partnerships.
Second, within this space, there are still several different ways in which partnerships can occur (M&A, Licensing, Strategic Alliance, etc.). By additionally observing these endpoints, we can further characterize the manner in which different types of startups are most willing to work with larger organizations.
Up to this point, we have been treating the choice of large companies in the database as fixed, but of course they are not. However, these sorts of randomized and targeted message experiments can easily apply to them as well. For instance, an open question is what do large organizations actually care about, and who are they willing to partner with? Are their stated goals consistent with these partnership choices? By randomly varying what information our member organizations see about our startups, we can disentangle competing theories of partner choice.
Finally, in creating this database, we are also given access to a large pallet of startups and partner organizations. We aim to leverage this access with the tools of experimental economics to gain insight into how these organizations make decisions intertemporally, under conditions of risk and uncertainty, and individually compared to in groups. Doing so will provide us with new insights relevant to economic and management theory.
Ultimately, the platform will not merely serve as a basis for research. Having completed our research experiments, and armed with better knowledge about the needs of startups and large organizations, we can then integrate such insights into the matching engine on the platform and make the platform available widely to large organizations who are seeking to find startups with specific technologies or solutions. CHIP will enable users from the TRISH, government (such as NASA), academia, and industry to connect with a global network of entrepreneurs and researchers. Investors and funding organizations can via CHIP engage with cutting-edge technologies being developed by innovators in universities, hospitals, research institutes, or industry startups.
The platform will be developed for us by Energizing Health, a Houston-based non-profit health innovation organization, which will later also host the platform for open use by healthcare organizations. Energizing Health works with a large national network of healthcare organizations such as Biden Cancer Initiative, MD Anderson Cancer Center, AARP, Humana, Kindred Healthcare, Merck Pharmaceuticals, athenahealth, Cambia Health Solutions, Ascension Ventures, the American Heart Association, a consortium of the nation’s top pediatric hospitals, and large brands like AT&T to solve health innovation dilemmas.
The PIs and Energizing Health have already begun to conduct a field test of certain assumptions around successful company selections for an event produced by Energizing Health for the November Scientific Sessions of the American Heart Association in Anaheim, California. These assumptions and results will be tested two or more times in the next 12 months at Energizing Health events in March at SXSW Interactive. One test will be for the Connect to End Cancer event produced by The Biden Foundation, MD Anderson Cancer Center, AT&T, and Merck. The other will be at Impact Pediatric Health led by the top eight pediatric hospitals in the United States.
Eventually, we intend for CHIP to be a unique one-stop-shop of collaboration, investment, and incubation for biomedical technologies. The CHIP searchable database will be populated from leading innovation hubs and sources (see below) as well as by all partner organizations using CHIP.
Customizable to meet the needs of the TRISH as its first user, CHIP v1.0 will uniquely tailor search parameters to align with NASA’s Human Research Program (HRP) risks, gaps, and concerns found at https://humanresearchroadmap.nasa.gov. Subsequent iterations of CHIP (see below) will be designed to meet the needs of partners in the broader healthcare community, including but not limited to hospital systems, payers, research universities, government agencies, angel investors, accelerators, nonprofit foundations, and more.
CHIP will first be populated with companies and technologies from the existing professional networks of TRISH Development Specialists. Following this, Energizing Health, our development and operations partner, will leverage its strong partnerships with health innovation stakeholders for future development and use of the platform. The value added to the TRISH in seeding this platform is the continual stream of quality opportunities which will be added to the platform by its growing network of partners. This will provide a platform by which the TRISH will not only identify disruptive health and performance innovations, but prospective partners who may match the TRISH’s investments and reduce the costs of advancing technology development.
Each technology entered into CHIP will be categorized and tagged by the following categories:
Risk(s) – according to NASA’s human research program roadmap
Gap(s) – according to NASA’s human research program roadmap
Stage – Pre-clinical, clinical, FDA approval, pre-sales, on market, etc
Type – digital health, medical device, biological, pharmaceutical, etc
Source – entrepreneur, company, academic, government, etc
Investors – current & previous
Advisors – current & previous
Other fields – as defined by TRISH super-users
CHIP’s database will be searchable by risks, gaps, innovators, academic groups, development stage, and more. Targeted and tailored notifications will be sent out to users when new technologies are added to classifications that they are interested in. Users can self-identify as an ‘innovator’ or ‘funder’ on the platform, in order to uniquely allow permissions such as data entry/edit, comment-only, or view-only access. Once the platform has over a thousand users, a privilege system will be implemented to enable users to “moderate” or “edit” entries.
CHIP’s searchable database will first be populated by TRISH personnel as well as curated data from leading innovation hubs, including but not limited to CrunchBase, AngelList, the National Institutes of Health (NIH) database of funded intellectual property from the National Library of Medicine, and others. CHIP’s versatile functionality integrates with leading workflow management tools to facilitate fast and simple collaboration across teams. CHIP is a stand-alone solution but will be interoperable with Slack for internal communication and smooth integration of new collaborators. Services including Amazon Web Services, Google Drive, Dropbox, and HIPAA-compliant Box, can be used on CHIP for file sharing, across all levels of IT security needs.
CHIP will have three initial versions, of which V1.0 will be covered by this grant:
- CHIP V1.0 – initial product and testing platform – basic features of data entry and reporting and searching TRISH’s proprietary database of startups as well as Application Program Interface (API) calls to current freely available databases of known startups such as CrunchBase and/or Angelist. (This proposal will only build and deliver CHIP V1.0).
- CHIP V2.0 – additional functionality for business users such as investors; enabling a request for proposals within CHIP. This will be pursued after the culmination of the grant activities, by Energizing Health.
- CHIP V3.0 – additional API access to commercial databases such as Pitchbook and CB Insights (that have a substantial annual fee to access associated) and a review of light artificial intelligence applications to normalize the data for user review. This will be pursued and financed by Energizing Health
CHIP V1.0 itself will have three phases:
- Phase 1: Build database, populate with TRISH data, customize user interface and reports, user test
- Phase 2: Recruit three partners to enter additional data, develop search algorithm, test reports
- Phase 3: Refine search algorithm; customize reports for TRISH, and test efficiency of identifying technologies suitable for TRISH funding, connect with an API to CrunchBase and/or AngeList and deliver back suggested healthcare startups.
CHIP V1.0 version will be entirely tag- (or “label”) based architecture. It will initially model and test users, companies, team members, and contributors, and IP (i.e. user Bob Jones is a reviewer for TRISH, Company is Spacely Sprockets, George Jetson is an employee at Spacely Sprockets, and Space University has licensed hand-held tricoder tech to Spacely Sprockets). Beyond that bare model, all attributes will be stored as simple tags. So “device_approved_by_FDA” is a tag that a company gets to have or not have, and so is “team_lacks_sales” etc. The tagging model is intentionally simplistic, it will allow us to understanding labeling and tracking priorities carefully before we enrich the data model to support more specific use cases. A comprehensive tagging model will be used that will work with any database of interest. Users will just search for the combinations of desired tags.
The CHIP knowledge base will expand through the investments by partner organizations with interest such as the Pediatric hospitals found at Impact Pediatric Health (see http://ImpactPediatricHealth.com), a collaboration of eight governing hospitals produced in partnership with Energizing Health that reflect a broader consortium of over fifty Pediatric hospitals curating technologies around shared technology and innovation needs. (See Letter of Interest at the end of this proposal.) Future partners such as these will share in accessing and increasing the content of the database and through this could potentially co-fund technology development with the TRISH.
Technical Development and Hosting
All software development and hosting will be completed by our implementation partner using appropriate languages and toolsets and integrating advances in computing and programming techniques as they emerge over the course of the grant period. Likely development languages and frameworks include Laraval and node.JS. Energizing Health will host CHIP on their website, while the actual code will reside and run under a cloud service, likely AWS.
Ongoing Management, Updates and Support
Energizing Health, our implementation partner, will manage the CHIP program after the end of the grant, managing and maintaining the software, providing periodic updates as needed and supporting users of the system. Energizing Health will first enter into a period of cost-sharing with its partner organizations to maintain the technical support staff necessary to provide ongoing updates, bug-fixes, training, and more to keep users successfully collaborating on CHIP. During this support and discovery period, system documentation and a version system of bug fixes will be implemented and tracked for two years at Energizing Health’s expense. After this period, future monetization of the CHIP system will pay for ongoing support and maintenance. If for any reason this monetization does not materialize, the source code and data for CHIP will be delivered to the TRISH.
Contributions of Project as Compare to Existing Databases
While various databases of startups exist today, they are siloed, difficult to merge, require myriad subscriptions and payments to access, and are not indexed or collated in a manner that provides for easy search for specific technologies. CHIP aims to solve this problem by cleaning, integrating and merging these multiple databases, using machine learning and other proprietary techniques to index, tag and curate a searchable database targeted towards healthcare-applicable technologies that TRISH and other healthcare users are searching for. This is no small undertaking and not something that other database providers are focused on or, based on our conversations with some, intending to focus on currently.
Deliverables and Key Performance Indicators for the Project
Once CHIP V1.0 is fully developed and validated, the PIs together with Energizing Health will work with the TRISH designated points of contact (the Development Specialists) to customize and refine CHIP for TRISH use. The deliverable will be a CHIP V1.0 knowledge base that will include up to 500 tagged and searchable entries selected by the TRISH Development Specialists. The platform will then be made available for the TRISH personnel to continue to populate and query for appropriate technologies. TRISH personnel will also receive ongoing customizable alerts when new entries of interest are added to the database.
Within six months of the conclusion of the research phase, the PIs will deliver a minimum of one research paper detailing their findings from tests on the platform which will include recommendations as to how the TRISH can maximize their strategy to engage startup companies in solving healthcare challenges for spaceflight.
Longer-term success, which will only be measurable some time after the grant period has concluded (with version 2.0 of CHIP developed by Energizing Health and integration of partner organizations who will utilize the data—such as TRISH), will be measured by the extent of match-making between startups and large organizations, user satisfaction in reducing the time and dollar costs associated with searching for specific startup technologies.
Note that since all information about the startups comes from existing datasources (public and proprietary through Energizing Health) and webscraping, no cooperation is required from startup companies. We also note, however, that engagement with the platform has an obvious return on investment to large organizations and startups alike. We expect the clear value proposition of being able to connect with potential customers and partners (startups) and being able to find needed technologies and startup partners (large organizations) should promote engagement with the platform. Startups do not need to know about the platform—they will be contacted by the large organization partners interested in working with them, through standard channels such as Linkedin and contact info on webpages. Traffic by large organizations will be driven by the web of partners who work with Energizing Health and email and digital marketing by Energizing Health as well as advertising at Healthcare related events and conferences attended by large organizations.
CHIP V1.0 Platform development is estimated to take approximately 6 months. An additional 12 months are requested for database population and expansion, algorithm development and testing. Six additional months are allocated for running hypothesis-based testing. A total of 24 months are requested for this project. Expenditures of the bulk of the grant money are anticipated to take place in the first year.
Wang, Yongzhi, and Nandini Rajagopalan. “Alliance capabilities: review and research agenda.” Journal of Management 41.1 (2015): 236-260.
Yael Hochberg, PhD, Rice University
Professor Hochberg’s research and teaching interests are focused on entrepreneurship, innovation, and the financing of entrepreneurial activity. Her research focuses on the venture capital industry, accelerators, networks and corporate governance and compensation policies. In addition to her doctorate in finance from Stanford, she holds a B.Sc. in IndusTRISHal Engineering and Management from the Technion-Israel Institute of Technology and an A.M. in Economics from Stanford University. Her research has been published in top tier journals, including Science Magazine, the Journal of Finance, the Review of Financial Studies, the Journal of Accounting Research, and the Journal of Financial Economics, and has been presented at numerous universities and governmental bodies around the world. Prof. Hochberg serves as the Head of the Entrepreneurship Initiative at Rice University and as Academic Director of the Rice Alliance for Technology and Entrepreneurship. She holds a Research Affiliate position with MIT’s Sloan School of Management and is a Research Associate at the National Bureau of Economic Research.
Prof. Hochberg is also Managing Director of the Seed Accelerator Rankings Project, which publishes the annual ranking of accelerator programs in the U.S. Prior to her appointment at Rice, Prof. Hochberg was previously on the (tenure track) faculty at the Kellogg School of Management at Northwestern University and the Johnson School of Management at Cornell University and most recently was a visiting faculty member at the MIT Sloan School of Management, where she retains a research affiliation. She serves on the advisory board and board of directors for a number of early stage startups. In 2015, she was named one of the world’s 40 under 40 best business school professors by Poets and Quants. In 2016, she was awarded the Ewing Marion Kauffman Prize Medal for Distinguished Research in Entrepreneurship.
Daniel Lee, PhD, Rice University
Dr. Lee is a postdoctoral fellow in Entrepreneurship and Innovation at Rice University’s Jones Graduate School of Business, and Research Associate at the Rice University Liu Idea Lab. He earned his PhD in Economics from Georgia State University, and holds his BA in Mathematics and Economics from Emory University. During his time at GSU, Dan was a CEAR Scholar and was the recipient of a World Institute for Development Economics Research grant from the United Nations University. He has published in outlets such as Science and the Journal of Economic Behavior and Organization. In addition to Entrepreneurship, his research interests include Experimental and Behavioral Economics and Applied Microeconomics.