By NEIL JENNINGS & VINCE KURAITIS

This entry is part 5 of 5 in the series Platforming Healthcare — The Long View

In previous posts in this series, we have covered the definitions of Unified Digital Health platforms and whether “EHRs can become UDHPs.” In this follow-on post, we’ll talk through the requirements for success for a UDHP and which types of healthcare organizations are best suited for which types of UDHPs. This post will build on findings from the previous posts.

The Market Needs UDHPs: Key Takeaways from Previous Posts

Key Takeaway 1: The healthcare industry needs UDHPs to create a centralized, common architecture for healthcare organizations

Key Takeaway 2: The healthcare organizations leveraging UDHPs will achieve a myriad of benefits, from competitive advantages to clinical, financial, and operational gains

Key Takeaway 3: UDHPs are not all-or-nothing or mutually exclusive from EHRs. As we explored in our last post, EHRs could expand into UDHPs. These EHRs as UDHPs (or the relative platform of platforms) may be the optimal choice for some market segments. EHRs may also be accommodated into cloud-first UHDPs.

Key Takeaway 4 / Guiding Criterion: This post will focus on US regional and local health systems and outpatient groups of all sizes.

The Approach: Market -> Segments -> Options -> Fit

• For this post, we will start from the top-down market perspective, analyzing the overall market landscape.
• Once we have described the landscape, we will call out the key segments (organization types, sizes, and profiles) that we will be evaluating.
• At this point, we will approximate IT budgets and IT team sizes by organization type to determine capabilities of building as opposed to depending on partners and vendors.
• Then, we’ll review the constraints for implementation and ownership, outlining the drivers of UDHP fit.
• Next, we’ll break down the different ways UDHPs can be developed and maintained.
  • Leveraging an EHR as UDHP
  • License from UDHP vendor
  • “Home grown” cloud-first solution
• Finally, we’ll crosswalk the segments and the optimal option for each segment, based on their specific needs and estimated IT and budgetary resources.

The Healthcare Market & Major Health System Segments

Starting with a compelling graphic from the Kaiser Family Foundation, we see a 2023 breakdown of the total US healthcare medical expenditure, totaling ~$4.9 trillion.

While the total healthcare spend that occurred in hospitals is an astounding ~$ 1.5 trillion, accounting for 31% of total healthcare spend, this leaves much of care outside the four walls of hospitals. This amount of care occurring outside of hospitals aligns with efforts to push patients into less acute care settings, emphasizing preventative, proactive medicine instead of acute, reactive medicine. As the need for UDHPs applies to more than inpatient hospitals, we will also review the other segments highlighted in the pie chart, including: outpatient clinics and practice groups, and “other health” containing services delivered at other contexts like PACs and SNFs, and Ambulatory surgical centers.

Table 1 below illustrates the key segments as we define them, including the rough count of total entities, their relative breakdown by size, estimates of annual revenue, and subsequent estimates of IT budget. We are estimating typical IT budgets as being between 2-5%, based on DefinitiveHealthcare and WittKieffer sources; we understand that a few organizations are higher or lower than this range. The objective, without spending too much time on the science project, is to rough out IT budgets by segment to align organizational resources and needs with the best-fitting UDHP.

Table 1

United States Healthcare Segments: IT Budget Analysis

1 (AHA – 1957 (unaffiliated) + 227 (small systems)), 2 (Statista), 3 (Futuremarketinsights), 4 (Businesswire), 5 (NIH), 6 (Chartis), 7 (DefinitiveHC), 8 (DefinitiveHC),9 (GrandviewResearch, revenue per MD) , 10 (AMGA), 11 (Grandviewresearch), 12 (Definitive Healthcare), 13 (Witt Kieffer), 14 (Precedence Research), 15 (Ibisworld as of 2023), 16 – these companies likely skew higher as they are remote first, depending on digital care platforms.

We are also including two succinct charts from the Witt Kieffer piece, because they tell a clear story about the huge variance in IT spent across the healthcare continuum.

Source: WittKieffer

Source: WittKieffer

Defining Fit: Segments (and their needs) Are Not Created Equal

Table 1 above highlights the wide range of healthcare segments that make up the multi-trillion dollar US healthcare market. As these organizations range from a few physicians to hundreds of hospitals spanning vast geographic regions, it should not surprise anyone that the organizational needs will vary greatly.

While the UDHP market is still in development, here are some of our considerations on what will define successful architectures and implementations. We summarize the determinants of organizational fit into two buckets: Capabilities & Needs

Organizational Capabilities

As shown in Table 1 above, IT budgets are finite, along with IT and technical teams’ time. Finding the right solution for the right organization ensures that wasted time and money are avoided. The UDHP should “fit” your organization’s capabilities.

Budget & Staff

Budget is an obvious starting point, with the 2-5% of annual revenues providing clear constraints. A small practice with a few hundred thousand dollars in annual IT budget should not be considering a multi-million dollar Epic install, of course.

Directly correlated with budget is the size of your IT team. There are staffing ratio discussions, with industry standard recommended ratios of IT staff to non-IT staff from 1:50 up to 1:150 or so. While somewhat obvious for the smaller organizations (if you have 5 clinicians and 3 support staff, you likely have a singular IT person who may also be an office manager), this is a consideration for the medium to large organizations. If your total team size is limited, adding new technologies to manage in-house can be daunting. Staff and team size come into play not just for longitudinal maintenance, but also for implementation.

Budgets can be used to determine which UDHP options are feasible, but staffing will influence how the UDHP will be implemented and maintained.

Current IT Footprint

The organization’s current technical footprint is also an important factor when determining the optimal UDHP implementation approach. As this is directly related to the IT strategy, we will also frame this section as some high-level guiding questions:

• What are the core systems of workflow and of record?
• Which of the following does the IT footprint encompass?
  • Software as a Service
  • Platforms as a service
  • Infrastructure as a service
• Does the organization build and maintain applications, portals, and systems from scratch?
• Are current IT systems hosted locally? Hybrid? On public clouds?
• Does the organization have a development team? A DevOps team?

Organizational Needs

The complement to capabilities is the organization’s current needs. Overbuilding or overinvesting is not the goal; the goal is finding the right fit. To define organizational needs, we will break them down into drivers of complexity, the need to stay competitive, and IT strategy.

Drivers of Organizational Complexity

While delivering healthcare in today’s digital-first world is complex for all providers and practitioners, organizations operate at different levels of complexity based on their scale, scope, business model, and footprint.

Starting with scale, larger organizations typically have more complicated needs. As the number of beds enters the thousands, the number of providers exceeds 2,000, and you start entering millions of annual visits, both the variety of interactions and the volume of interactions increase. In addition to the number of people (patients, providers), the number of physical locations is another major driver of complexity, where a single, standalone small hospital may be simpler than a multi-state, multi-specialty physician group.

The scope of the organization’s care services is another major factor. Are there only outpatient/ambulatory practices? Are they only primary medicine? Or, are there specialty groups? Is there a mix of acuity settings? Does the organization have urgent care, emergency medicine, home health, hospital outpatient specialties? Does the organization have its own labs, imaging centers, and surgery centers? As you continue to add more types of care that can be delivered, the IT infrastructure is challenged in new and exciting ways.

Provider organizations’ business models also play a key role in organizational complexity, with the shift towards value-based care and risk-bearing models. As provider organizations shift towards risk-bearing models, it is often an incremental shift with risk-based and fee-for-service models running in parallel. In addition to the (relatively) simpler configuration of adding another insurance plan, the data and care delivery coordination of risk-based plans are an order of magnitude more complicated than traditional fee-for-service care models. This shift has expanded beyond enterprise health systems and large hospitals and is cascading through large ambulatory practices, which have distributed footprints and historically smaller IT budgets.

Geographic footprint is a major driver of organizational complexity and IT (among other) needs. Is the organization a singular location or spread out within a city? Is it distributed across multiple states, countries (Canadian presence), or other geos (tribal nations, US territories, etc.)? As organizations expand to multi-state footprints, for example, there are differing regulations for licensure (the scope of NPs for example), e-prescribing, and other localizations like regional language needs: this all must be accommodated in EHRs and other systems. This is all on top of the need for referral networks, reference labs, pharmacies, and local health information exchanges. In addition, there are regional privacy, data sharing, and other requirements that apply to cross-state or multi-region businesses. When geographies span multiple types of borders, it gets even more complicated.

Geography is more than just technicalities related to data sharing, licensure, and care delivery: we cannot forget the people and their cultural and regional differences. The US is an incredibly diverse country where there are 430 languages spoken across the states and territories. Addressing cultural, language, and regional differences is a complexity that many large health systems and national healthcare providers face, and there are a myriad of studies showing that patients are more engaged, have better outcomes, and are more satisfied with care delivered in their preferred language. Anyone who has worked on an internationalization or localization effort for a product will understand the complexity, but the implications of languages, regional differences and norms, are far reaching and cannot be underestimated. They expand beyond multi-language support, impacting content (care summaries, letters, patient instructions), and even office schedules.

Organization Needs to Remain Competitive / Differentiated

Beyond just managing their own complexities, healthcare organizations are typically run as businesses (there are some not-for-profit organizations, including a national leader that we highlight later). With the shift to lower acuity care contexts, private equity rollups, regulatory changes, and the constant need for more patient-centric care – the market and competitors are changing quickly!

Organizations have rapidly changing needs in order to stay competitive and differentiated in the digital-first healthcare ecosystem. Both organizational needs and patient/provider expectations are driving initiatives for lasting differentiation.

Starting with organizational needs, some organizations are building competitive data moats, leveraging differentiated data access and repositories, to remain competitive. Whether accessed as part of a digital network or exchange or owned wholly internally, this approach relies on IT infrastructure to securely store, process, secure, and broker data to connected systems.

In terms of expectations, while patients today have more control of their own healthcare records, it is still not enough. Leading organizations realize that patients do not want to interact with separate care contexts, specialties, and facilities individually. Instead, they are prioritizing unified patient experiences (a component of UDHPs) to simplify patient interactions across their technical and operational ecosystem. This transition to patient consumerism is providing another point where organizations compete: patient engagement and loyalty.

IT Strategy

As organizations get larger in size, they have more defined IT strategies, with enterprise organizations maintaining multi-year roadmaps. These roadmaps outline the prescribed approach to achieve organizational priorities and define the projects and budgets for upcoming years.

While this point is foundationally important, it also varies greatly, depending on the idiosyncratic needs of each organization. However, there are some common themes and decisions we can typically discern when reviewing a roadmap. We’ll frame them as questions below:

• Is the organization trying to expand or shrink the scope of its IT function?
• Is there a strong preference for building vs. buying vs. partnering?
• What key unmet needs have yet to be addressed and are prioritized?
• What is working today (and notably excluded from roadmaps)?
• Where does organizational leadership have their sights focused to increase efficiency, competitiveness, differentiation, and care quality via IT efforts?

Defining UDHP Options for Different Healthcare Segments

While healthcare at-large needs UDHPs, the specific needs vary by organization size and complexity. Related, not all implementations of UDHPs will need to be the same, and we see three primary paths for organizations to implement and architect their UDHPs.

Mapping UDHP Options to Market Segments

Option 1: Leveraging an EHR as a UDHP

There are several drivers that push healthcare organizations toward their EHR as the default UDHP. These systems already accommodate the complexity, sensitivity, and technical requirements for operating in a regulated industry. Some of the major tailwinds, as covered in our previous post (“Can EHRs be Health System’s “Platform of Platforms” (UDHPS)) are as follows:

• EHRs Currently Own the Customer Relationship
• Many Customers Have an “EHR-First” Preference for New Applications
• Epic and Oracle Health are Making Strong Movements Toward Becoming UDHPs

Due to budget constraints, the EHR may be the only practical UDHP option for many organizations. Smaller practices and medical groups may have a single IT person, responsible for EHRs, hardware, software, printers, fax machines, and everything else. If IT budgets are starting around $100,000 per year, and EHRs may take up to half of that, there isn’t budgetary or staff room for other systems. Elation calls out that the expected license cost of an EHR per clinician is around $1,200 per year; Folio3 shares a similar number. This is before maintenance, support, training, and other expenses related to the EHR, and before other IT-related expenses not related to the EHR. When the EHR makes up a substantial portion of the budget, it makes sense to look there first.

In addition to those points mentioned in the previous post, when EHRs are tailored to smaller organizations, vendors understand that their customers need a partner that fills in gaps. athenaHealth’s athenaOne package offering, for example, includes add-ons that complement the core EHR across the organization: from patient engagement and communication, to pre-visit and sched/reg support, to RCM and claim submission support. Positioned as an “all-in-one practice solution,” they are framing the offering as more than an EHR and offer many UDHP benefits.

Moving upmarket to larger physician groups and small hospitals or health systems, IT budgets get larger as revenue grows and care contexts get more complicated. As we begin to see multi-state footprints, large multi-specialty groups, and acute care settings, organizations may outgrow the ambulatory-focused EHRs like athenaHealth, ECW, and others. This is where Epic and Cerner step in and are happy to meet these organizations’ needs.

Epic’s Connect (formerly known as Community Connect), for example, is a specific offering that enables large health systems to sublicense their Epic instance, provisioning access to regional hospitals or health systems. While this option allows smaller healthcare organizations to leverage the benefits of Epic, the cost can be an order of magnitude higher than outpatient, license-based EHRs.

In 2024, OhioHealth Van Vert Hospital invested $12MM to transition to Epic via CareConnect. Leveraging the same budgetary assumptions as above, if the EHR is ~50% of budget, and budget is ~2-5% of revenue, this suggests that the hospital has >$480MM in annual revenue ($12M * 2 / .05). Implementing Epic as a standalone health system or hospital group is expensive, and a 2024 post from Becker’s notes that costs may start in the tens of millions, but can reach a billion dollars or more for larger health systems.

As the IT budgets get larger, technical teams increase in size and capabilities, and it naturally follows that more advanced options are explored or required.

Option 2: License from UDHP Vendor

As denoted above, not all organizations have sufficient IT budgets and staff to architect, build, and manage their own system. But, for large health systems with more than a billion dollars in annual revenue (estimated at ~175 or so in the United States), these options start to become feasible if not necessary. These large hospitals and health systems may begin looking beyond their EHR to meet their UDHP and platform-of-platforms needs. Why?

As highlighted in the previous post:

• EHRs Carry a Lot of Baggage
• Customers are Skeptical
• EHR Analytics Are NOT Optimized To Achieve Critical Health System Objectives
• EHR Switching Costs are Diminishing
• Cloud Native Platforms Accelerate Innovation and Performance
• It’s Not in EHR DNA to Become A Broad-Based Platform

The last post also shared a HIMSS study that showed that the average hospital had 16 EHRs in 2016. While this number has likely consolidated down in the past decade, multi-EHR and multi-system-of-workflow architectures are very much a reality today. Focusing on the complexity angle in this piece, we will highlight four additional, complementary points as to why an outside-the-EHR UDHP can be the ideal solution for larger hospitals and enterprise health systems.

There are four primary reasons that we will highlight the need for a UDHP to exist beyond the core EHR:

1. No singular EHR can accommodate everything that health systems need to do
2. Some point solutions are still better than their EHR-native competitor
3. Some organizations leverage multiple core EHRs (there are organizations that use both Cerner and Epic across their hospitals)
   1. Whether due to M&As, IT or clinician preference, or other reasons, some health systems still use a mixed EHR footprint.
   2. This presents in different ways, but is typically one of the following:
      1. One system for inpatient, one for outpatient
      2. One system for clinicals, one for RCM
      3. Or, distinct systems at distinct locations (often due to M&A)
4. M&As and rollups are still running rampant
   1. Mergers and acquisitions combine more than market footprints and clinical capabilities; they also bring IT teams and infrastructure. This is where we begin to see the real glimpses of Frankstein’s monster.
   2. Some stats from Kaufman Hall:
      1. 72 total transactions in 2024, with 27 not including a divestiture. While this number may not seem that high in a vacuum, the total picture of the last decade+ shines light onto the real volume of these transactions. From 2014 to 2024, there were 933 total M&A transactions. That is a lot of technical complexity to clean up.

In terms of licensing an outside-the-EHR platform, the first post in the UDHP series includes a comprehensive list of possible vendors, building from Gartner’s original list. We have also highlighted in previous posts that the license approach is not an all-or-nothing approach, and typically involves a hybrid model, building on top of one or several EHRs. Each organization’s specific needs are different, and they may use their licensed platform from a UDHP partner to varying degrees.

Option 3: “Home-grown” cloud-first solution

The final option that we’ll cover in this post is for the leading health systems, and likely limited to the top 15-20 systems due to complexity, expense, and staffing requirements. Progressing beyond the cloud and platform vendor partnerships mentioned above, these systems are building enterprise data platforms and UDHPs directly within cloud environments, leveraging cloud services from AWS, GCP, Azure, InterSystems, or others.

Starting with a mini case study. We plan to further expound upon Ascension’s work in a future post.

The Situation: Ascension checks all of the boxes for a large, complex organization.

Ascension’s Scale: 16 States, 94 wholly owned or consolidated hospitals, 27 additional hospitals through partnerships. 99,000 associates, 23,000 providers.

Ascension’s Scope: Inpatient, outpatient, and everything in between.

Ascension’s Staff: ~3,000 working in IT (per LinkedIn)

Ascension’s IT Footprint: Heterogenous, 18 EHRs, 40+ patient portals

Need for differentiation / competitiveness: High, both regionally and across the 16 states

The Problem:

• Patients are not aware of “care contexts,” they expect a unified experience.
• Ascension’s diverse set of EHRs and myriad of other systems were not tightly integrated
• No singular EHR nor operational system was a complete source of truth
• There were data variations across systems that compound and exacerbate experience inconsistencies across consumer digital touchpoints

The Solution:

• Ascension created an “Enterprise Data Service Layer (EDSL),” mapping data from multiple EHRs into a unified, centralized FHIR schema to facilitate frictionless consumer access to normalized and harmonized data
  • This relies on deep partnerships with EHR vendors and other partners.
  • FHIR, not fully adopted by all vendor partners, was made more accessible using REST APIs and a GraphQL orchestration layer to provide data on demand to digital applications
• A unified patient experience, “AscensionOne” was developed on top of the EDSL to provide a singular place for patients to book visits, get care, manage their health records, and pay for care.

The Outcomes So Far:

After one year, over 400,000 members are using the unified patient experience, built on top of the Enterprise Data Service Layer. The application is averaging 4.8 stars in the app store.

Thoughts from a leader on EDSL efforts:

MJ Barrois, Senior Director of Product Management:

“Our mission to better serve our communities was a key driving factor behind our investment in building our internal platform. EDSL represents our commitment to help shape a future where highly interoperable and portable data is readily accessible to all of our patients. To provide care to those most vulnerable and in need, we strive to be good stewards of their data and we work closely with our EHR partners to make healthcare data securely available, easily discoverable, and understandable to the patients we serve.”

Summarizing the home grown option:

While this option is not the optimal choice for all organizations, Ascension is paving a path that will be easier for others to follow. As we revisit the maturity curve shown above, the technical capabilities and requirements will get easier year after year. AI is accelerating data mapping and infrastructure optimization. We expect that the “homegrown” solution will be more attainable for the top 50 or so health systems in the next few years. If there is ever any doubt about what the cloud systems can do, we’d suggest checking out this cloud comparison cheat sheet from ByteByteGo. You can quickly see that the cloud platforms can handle pretty much anything you can throw at them if you have the right team and budgets in place.

Looking ahead

Today, AI and our better-than-ever hardware continue reducing barriers to entry, and the need for UDHPs will rise as organizations increase their IT footprints. Choosing a platform to build around is the important decision organizations should make today. This platform decision will be an integral component of organizations’ AI strategies, which is an essential lever for short term competition and long-term differentiation. There is a lot of noise around AI implementation best practices today, but the consensus (per AIDoc) appears to be that a centralized, governed platform approach is best, which relies on unified digital health platforms and data.

As we look a few years out, a highly possible scenario is that by 2030, cloud-first implementations like Ascension’s will be commonplace not just among the top 7 health systems, but the top 50. Hosting EHRs (like Epic and Cerner) on public clouds, pioneered by Geisinger, has been de-risked. Once in the cloud, it’s only a matter of time before the other cloud-first capabilities are evaluated and deemed a step forward from legacy, healthcare-specific platforms and services.

For the next cohort down in size in the same period, it is highly likely that EHR-complementing UDHPs will be licensed by the next top 150 health systems and hospitals. These already cloud-first systems will play nicely with cloud-hosted EHR systems and provide access to cloud benefits not attainable due to staff and budgetary limitations of these smaller organizations.

What is the final frontier? Ideally, by 2030, we will also begin seeing uptake in openEHR and open-syntax, cloud-first platforms in the United States. Epic isn’t going anywhere, yet.

Vince Kuraitis JD/MBA is an independent consultant with over 35 years’ experience across 150+ healthcare & tech companies. Neil Jennings is a solutions engineer on the health care team at Akamai Technologies. Vince publishes The Healthcare Platform Blog, where this post first appeared. The authors thank several colleagues for their review and commentary on an earlier draft of this post: M.J. Barrois, Senior Director of Product Management at Ascension; Jim Sabogal, Healthcare Vertical Lead at CDW; and Taylor Gartley, Director at Fold Health.