Data interoperability and the governance of public value

Introduction

In the face of increasing platform centralization and market power states wield interoperability as a tool to reconfigure platform markets. Broadly, interoperability describes the capacity for different technical systems to operate together (Kerber and Schweitzer, 2017). Using a credit card across international borders, making a phone call to a friend on a different telephone network, or driving a train across the country on rails owned and managed by different organizations each requires interoperability. This commentary explores a set of U.S. regulations that seek to advance data interoperability within healthcare markets. Via this case study, I demonstrate how U.S. policymakers seek to wield data interoperability and platformization, not only to reconfigure power relations between powerful market actors, but to align healthcaremarkets with the production of public value. I argue that data interoperability is an insufficient tool for structuring and ensuring the production of public value from data.

Data interoperability is an infrastructural precondition for the production of value in the platform economy. Providing third party developers or “complementors” (McIntyre and Srnivasan, 2017) with access to standardized, controlled data streams is central to the way that platforms produce value on top of the data flows that they structure and intermediate. This data interoperability is achieved primarily through “application programming interfaces,” or APIs. Through APIs, companies like Google, Facebook, and Amazon give third parties access to a controlled dataset about its users (Gawer, 2011; McIntyre and Srinivasan, 2017), enabling those third parties to build products and services that ultimately contribute to the market dominance and value of the platform itself (Hein et al., 2020). APIs enable data interoperability at speed and scale, thereby enabling data that is captured and structured by one platform to be automatically transferred to another platform for secondary- and tertiary-value production processes. In this sense, there is no platform economy without robust, widespread data interoperability.

Platforms exert immense control over the ways that data is structured and shared, and therefore over the kinds of value that can be produced from this data. Platforms like Facebook, Twitter, Google, and Amazon carefully control the design and implementation of their APIs in ways that benefit the platform company themselves and put third party complementors in a position of dependence upon the platform. When platforms decide that sharing data with third parties is not producing sufficient value for the company, they can make unilateral changes to what data is shared and how. For instance, in 2023, the social media platform Reddit decided that it was not gaining sufficient financial benefit from the products and services offered by third party complementors, and decided to drastically increase the costs of API access. Reddit make access to its APIs so expensive that it effectively shuttered an ecosystem of beloved secondary products overnight (Grantham-Philips, 2023). Twitter's decision to end free access to its APIs caused similar upset among communities and users (Hendrix, 2023). Through APIs, platforms exert unilateral control over the ability of third-parties to access and produce value from platform data.

While platforms wield significant market control via their control over APIs, platforms may also wield “adversarial interoperability” (Doctorow, 2023) to disrupt and reconfigure market power. Cory Doctorow illustrates how Facebook was able to interrupt MySpace's market dominance by scraping its user data and importing it to Facebook, enabling users to more easily disintermediate MySpace and adopt a new social platform. Wary of this potential for disruption via adversarial interoperability, Big Tech platforms create tightly controlled “data enclaves” (Birch, 2023). Through the control of these data enclaves platform companies can box out potential competitors through “lock in” effects. Katarina Pistor describes this dynamic of the platform economy, wherein certain Big Tech platforms wield their control over data enclaves in a way that doesn’t allow for meaningful competition “the end of markets” (Pistor, 2020).

Just as platforms can wield interoperability to structure secondary markets in ways that produce value for the platform themselves, states can enforce data interoperability to break up these data enclaves and reconfigure platform markets. This represents a form of platform governance undertaken at a protological or infrastructural level (Cohen, 2020). Morton et al. describe interoperability as a “supertool” of platform governance, enabling “light touch” regulation, in as much as this regulation encourages competition without directly dictating modes of value production (Morton et al., 2023).

The 2022 European Digital Markets Act is an example of what it looks like for a state to wield interoperability to mitigate the formation of data enclaves and platforms' monopolistic control over markets. The DMA outlined new interoperability requirements for messaging platforms like Telegram, iMessage, WhatsApp, etc. These requirements force the messaging platforms to relinquish some of the hold that they have over customers by requiring these platforms to let customers message across these platforms—much like any email client can be used to send and receive messages from any other email client. This intervention is designed to enable greater consumer choice across competing platforms, reducing consumer lock-in (Brown, 2020).

However, interoperability regulations are not just a tool for mitigating the monopolistic tendencies of data-hoarding platforms: they are also central to market-making. Data interoperability can be used as a tool of governance to push legacy markets away from value production based on data siloing and consumer lock-in toward value production based on data sharing across platform ecosystems. For instance, the 2010 EU Commission Digital Agenda highlighted interoperability as one of the most important barriers to value creation via digitalization (Kerber and Schweitzer, 2017). Likewise, emerging regulations in the United States seek to increase health data interoperability are designed not only to disrupt data enclaves created by dominant market players, but to spur the creation of platform-like markets within the healthcare industry.

In what follows, I explore how U.S. policymakers articulate platformization—enabled via interoperable data infrastructures—as the key to producing both public and private value. I then reflect more broadly on the role of interoperability as a device for governing the kinds of value produced through data.

Infrastructuring public value from data: platformizing the US healthcare industry

Data sharing in the U.S. healthcare system today is often constituted by fragmented, manual, nonstandardized data sharing across different, often competing, provider, and insurer organizations. If you want to see a specialist, your primary care doctor may fax your entire health record to that specialist's office. Your specialist may enter information from your visit into their proprietary health record system, then fax a PDF back to your primary care doctor to be entered manually into a separate, proprietary health record system. Alternately, your primary care doctor and your specialist may have a data-sharing agreement in place and a point-to-point technical interface enabling data sharing between their health record systems, meaning that data can be exchanged without requiring manual faxing and scanning by an office clerk. Or you may be so lucky that your provider and specialist have an even higher level of interoperability, particularly if they are in the same network or hospital system.

Without data interoperability, health information exchange is costly, slow, and the usability of information within healthcare workflows is significantly constrained. Insufficient health data interoperability problem has been the subject of increasing regulatory intervention at the federal and state level for over 15 years. In 2009, in conjunction with the Obama administration's landmark Affordable Care Act and as part of an immense economic stimulus package, the U.S. federal government designated around 20 billion dollars incentivizing hospital and provider systems to shift from paper-based records to electronic health records (EHRs) through the HITECH Act. This digitization of health records was articulated by regulators as part of the solution to an “inefficient” healthcare market, characterized by higher costs and lower quality of outcomes relative to other OECD countries. Digitization was promoted as a pathway optimizes the productivity, innovation, and social benefits of the healthcare market in the United States.

Yet even in the early phases of this monumental transition to electronic health records, consultants and policy makers argued that creating an efficient, competitive healthcare market would require more than digitalization of health records: it necessitated the intentional cultivation of health data interoperability (Thune et al., 2015). Subsequently, in 2016, President Obama signed into law the 21st Century Cures Act, which included provisions for increasing health data interoperability. These provisions included the creation of a national health data sharing network and increased requirements for EHR companies to ensure that health data could be “accessed, exchanged, and used” via APIs (21st Century Cures Act, 2016).

These interoperability regulations were motivated, in part, by the recognition that health data enclaves—strictly controlled by EHR companies, powerful medical systems and insurers—are not the best way to ensure the use of data to produce public value. A 2014 report commissioned by the federal government outlined the case for interoperable data infrastructures and laid the groundwork for Cures Act interoperability regulations. Much of the focus was on ensuring that EHR companies enable data exchange via APIs: “Health data in EHRs that support robust APIs are a more valuable public good than health data in siloed EHRs” (Data for Individual Health, 2014: 51). The report's authors frame health data “siloed” within EHRs as a latent source of public value, waiting to be tapped via interoperable data infrastructures. Subsequent health data interoperability regulations can be understood as attempting to shift the market away from a model of value-from-data predicated on data enclaving and consumer lock-in (Pistor, 2020) toward a model of value-from-data predicated on an ecosystem of products and services competing on their ability to use data in the most innovative, cost-effective and user-friendly ways.

In the United States one privately held EHR company holds 35% of the market share among hospitals (Joseph, 2024). This company holds significant market power primarily by virtue of its control over immense stores of health data, demonstrating robust network effects and “lock in” dynamics, wherein health systems are incentivized to join EPIC because it is the most centralized way to access and share patient information. Yet health data interoperability regulations aren’t designed to disrupt the EPIC's market power, but to move the industry toward platform business models, through which platform data is made available for secondary and tertiary forms of value production. The same 2014 report which helped to lay the foundation for interoperability regulations argues that shifting from “a small number of proprietary systems to a software ecosystem with a diversity of products and applications” would help create a more functional, efficient, and cost-effective healthcare market as patients and physicians “gravitate toward user interface applications that provide the best functionality and convenience” (Data for Individual Health, 2014: 48).

In short, health data interoperability policies were undertaken as a means to achieve platformization, itself seen as the best way to achieve both private and public value. There is much to scrutinize about the state's assumptions about how interoperable data infrastructures—and the platform markets they enable—will produce public value. This theory of change relies on two key assumptions: first, it assumes that health data exists in a “raw” form within EHR systems, readily repurposed and mobilized for the public good by secondary apps and services. Second, it assumes that there will be a higher degree of competition among these secondary apps and services than exists within the market currently, enabling “consumer choice” to drive forward the apps and services which most directly produce public value, in the form of improved care and patient and provider experience.

In the next section, I will unpack these two assumptions further, highlighting the limitations of data interoperability as a means to achieve public value.

Discussion

U.S. health data interoperability regulations are intended to enhance the production of public value from health data through shifting towards a platform-like market, wherein data is made available at speed and scale to secondary apps and services. Yet this “theory of change”, which sees health data interoperability as enabling increased production of public value from health data, relies on two key assumptions: first, that data exists in a “raw” form that can be readily taken up and mobilized for the public good; and second, that these new platform market will be more competitive, thereby driving the production of products which advance the public good (e.g. improved care outcomes, improved usability for patients and providers).

These assumptions are important to interrogate because they are common across other models of platform governance through interoperability.

Conclusion

Interoperability is articulated as a crucial tool for intervening in increasingly centralized and monopolistic platform markets. When wielded by states or platforms, interoperability can have an adversarial function, disrupting established data enclaves and reconfiguring market power. Yet, just as antitrust regulations alone cannot ensure that markets help to advance the public good (Bietti, 2022; Manne and Wright, 2011), data interoperability alone cannot ensure that platform markets are aligned with the production of public goods. The promise of “light touch” regulation via data interoperability in fact reflects a set of assumptions about how data is made to produce public value that deserve deeper scrutiny.