Tag Archives: AWS

Signify Premium Insight: Big Tech in Bed with Big EHR on the Cloud

This Insight is part of the Signify Premium Insights (SPI)-Digital Health service, which will launch on 9 January 2023. From that date, this and all SPI-Digital Health Insights will be available only by paid subscription. Click here for a free one month trial of this service. 

It is 2035. The last US hospital has finally migrated its EHR to the public cloud. It has been some digital transformation journey, spanning nearly two decades. In that time, hospitals have realised lower operating costs, lower total cost of ownership and the goal of seamless, system-wide information sharing. Big EHR vendors like Epic and Dedalus have answered customer needs. And Big Tech has played a central role in the transformation. 

Rewind to 2023 and public cloud providers Google Cloud, Amazon Web Services (AWS) and Microsoft Azure stand, at last, on the cusp of cracking the hospital IT public cloud market. Their success to date has been mixed: only around 15% of US hospitals had fully migrated to public cloud as of end-2022. But recent EHR hosting agreements between Epic and Google Cloud (in the US), and Dedalus and AWS (in Europe) promise to accelerate migration.  

However, on the notoriously slow-moving healthcare ship, and for many reasons, full migration will take more than a decade to achieve. 

The Signify View 

Epic’s November 2022 announcement that hospitals running its EHR would be able to run their EHR workloads on Google Cloud was a pivotal moment for the tech giant. Around one third of US hospitals run Epic’s EHR solutions. It is by far the largest health IT vendor (in terms of the number of hospital customers), and with a steadily growing market share. Google Cloud can reasonably hope that, eventually, it will host most (if not all) of Epic’s (growing base of) customer EHRs. Epic will pay Google Cloud for every gigabyte of storage and megahertz of processing power its customers’ EHRs need. With millions of patient records with Epic, that is a huge prospect for Google Cloud. But public cloud migration is lengthy and complex. Any migration benefits will take years to materialise.  

Short-Term Shot 

Shorter term, the Epic tie-up helps Google Cloud make up ground on AWS (the leading primary and secondary cloud provider to healthcare IT vendors) and Azure in terms of public cloud market share. Until now, Google Cloud has typically been used as a secondary provider to plug functionality gaps (health IT vendors do use different cloud providers based on functionality needs and their want to accommodate the cloud preferences of the payer/providers’ customers). Google is also running a clinical data search and summarisation project for MEDITECH’s Expanse EHR. 

Opportunities at Cerner now appear much more limited for Google Cloud at this time, however. Oracle’s 2022 acquisition of Cerner (and Oracle’s stated ambition to reduce Cerner’s reliance on Oracle’s competitors) will weaken Cerner’s relationship with AWS, which was implementing public cloud for Cerner’s customers. This will be to Oracle Cloud, not Google Cloud’s, advantage.  

Relationship dynamics are important, and Google Cloud will have to bear in mind its own patchy record in this respect, most notably with Epic itself. In 2020, the number one vendor told customers it would no longer pursue further integration with Google Cloud. It temporarily moved to AWS and Azure to provide cloud technology for medical research, data storage and file sharing. Google Cloud will see the latest tie-up with Epic as being both a long-term business opportunity and a chance to rebuild bridges. 

Benefits and Barriers 

Google Cloud says by hosting Epic’s EHR on its public cloud, hospital systems will have access to analytics and AI to improve patient outcomes. EHR vendors, meanwhile, see cloud as a way for their customers to reduce operating costs, total cost of ownership, enhance their products and improve the security of their data. 

Cloud is cheaper (in terms of total cost of ownership) and quicker to implement for providers than on-premise server rooms/racks etc, and ongoing maintenance costs are minimal. Cloud users simply pay a monthly fee. Cloud also has the obvious advantage of multiple sites being able to access the same EHR instance. Furthermore, hospitals need not employ IT staff (or consultants) to maintain their servers. Cloud offers far more predictable budgeting. Cloud advocates also say scaling is easier and that hospitals can channel resources towards care quality and patient outcomes rather than managing on-premise servers.  

The reality, however, is that most hospitals have invested heavily in on-premise data storage and server rooms, including hardware and software licences, in the relatively recent past. They have factored annual maintenance and infrastructure support costs into their capex and opex models. Despite the clear benefits of public cloud migration, some hospitals are understandably keen to maximise the shelf life of costly IT installations. Migration costs and disruption, and the development of additional network infrastructure, are also significant barriers in terms of transition. For some hospitals, it could take a decade or more to make the switch. In some cases, in the early stages of this process, hospitals are migrating to private cloud installations run by Epic and Cerner hosting their own EHRs, which are seen as more secure. A hybrid cloud model is also a potential intermediate step.  

There are justified concerns over breaches of highly sensitive Personal Health Information (PHI). Such incidences have previously strained relationships between EHR vendors and Big Tech. However, public cloud vendors invest billions each year on data centre security, many times more than a healthcare provider. Consequently, public cloud provides a more secure environment than on-premise.  

Off-premise EHR also raises the prospect of data latency, especially with regards to the storage of memory-intensive documents. And, finally, there are regulatory hurdles to public cloud migration. Some countries prohibit hospitals from storing patient information on the cloud. Others have data sovereignty laws which prohibit patient data being stored outside the country. This is an issue in regions such as the EU. 

However, changes in IT procurement and provider organisational structures, along with the advantages outlined earlier, are expected to counter these barriers. The growth of hospital networks and IDNs in the US, and regional procurement organisations in international markets (e g GHTs in France, ICSs in the UK and ACO-type health clusters in Saudi Arabia), mean that IT procurement and decision making is becoming increasingly centralised (around CIOs/CEOs or regional governments), and adoption planned from a healthcare network perspective, instead of an individual hospital perspective. This will have a positive impact in bringing cloud adoption to scale. 

Coming of Age 

A new era is dawning for hospital IT. The power struggles between EHR vendors and the tech giants which have flared up in recent years are being replaced by a more collaborative approach. This is illustrated not only in this month’s Epic-Google Cloud tie-up, but also an agreement last month between AWS and Dedalus which will see Dedalus customers access a range of cloud services including disaster recovery and improved security. Currently, AWS hosts Dedalus EHR solutions for more than 50 hospitals in the UK, with more migrations planned next year and in 2024. 

These partnerships promise to accelerate public cloud adoption, reducing costs and improving patient outcomes for hospitals, and opening up new markets and revenue pathways for the likes of Google Cloud and Epic, and AWS and Dedalus. The public cloud providers will become an increasingly important conduit for IT vendors delivering software to provider and payer organisations. 

Public cloud EHR migration is, however, complex and time intensive. It will require a change of mindsets before the benefits truly kick in across the public cloud ecosystem. 

Signify Premium Insight: Amazon in Prime Position with HealthLake Plans

This Insight is part of your subscription to Signify Premium Insights – Medical Imaging. The content is only available to companies that have subscribed to this paid-for service. To view other recent Premium Insights that are part of the service please click here.

Last week Amazon made clear its intentions in medical imaging, announcing two new capabilities in HealthLake focused on medical imaging and analytics.

The Seattle-based tech firm says that the abundance of data created in medical imaging is slowing down decision-making in hospitals. In response, the cloud vendor has launched Amazon HealthLake Imaging, which is designed to expedite medical imaging retrieval in clinical workflows, as well as powering existing medical viewers and analysis applications. This, the vendor claims, can result in considerable cost savings.

However, with Microsoft’s Nuanced-derived healthcare expertise, Google’s recent moves into medical imaging, and Oracle’s inherited incumbency via its Cerner acquisition, has Amazon done enough to win the interest of providers?

The Signify View

Tech giants’ interest in healthcare is nothing new. Amazon, like other Nasdaq darlings, has made various approaches to different healthcare markets over recent years, from the launch, and subsequent shuttering of Amazon Care, a primary care service, to its Amazon Pharmacy play. Recently however, several of the world’s biggest tech firms have redoubled their focus, setting medical imaging firmly in their sights. After Microsoft’s acquisition of Nuance, which closed in April 2022 and Google’s recent launch of its Medical Imaging Suite, Amazon has become the latest tech firm to make a concerted imaging effort.

Like Google’s launch before it, the launch of Amazon AWS’ HealthLake Imaging suite is not festooned with brand-new, never-before-seen capability. Instead, many of the tools and partnerships included in the package have been available previously in various guises. However, the new packaging highlights Amazon’s increasing focus on selling cloud services to acute and outpatient providers as interest in, and understanding of the technology increases. While many of the tools have been previously available, it would likely have taken an already knowledgeable user at a provider to capitalise and work out how best the range of tools offered by Amazon could be applied to their own imaging departments. The packaging and positioning of Amazon’s latest effort, however, should help providers more clearly appraise the potential of cloud adoption for their imaging departments, easing the transition for more mainstream providers.

Such positioning, however, is only enough to make AWS more accessible. What Amazon hopes will encourage providers to commit, is its boasts about price. In its blog detailing the new solution, Amazon estimates that HealthLake Imaging helps reduce the total cost of imaging storage by as much as 40%.

The Cost of Delivery

This figure, as is often the case with those used in marketing materials, should be taken with a pinch of salt. No doubt Google, Microsoft and other cloud providers harbour some technologies which are also designed to help reduce the cost of storing images on the cloud. However, the fact that Amazon has publicly stated the savings that providers can expect indicates the vendor’s confidence in its ability to offer providers an affordable option.

Cloud adoption, can, after all be stymied by the cost, or at least perceived cost, of making the transition. While this is less of an issue for flagship academic providers and the premium they are willing to pay to have the latest and most experimental technology, for the acute and outpatient providers, cost is a far greater barrier. If Amazon is to truly capitalise on the revenue-making potential that cloud provision in medical imaging offers, however, this mass market is ultimately where the vendor must target.

By highlighting the cost-savings providers can expect to make if they adopted Amazon’s imaging cloud solution, even if the actual savings delivered are not quite at the quoted 40%, the vendor hopes to overcome the perception that cloud is prohibitively expensive, and at least engage mainstream providers in a conversation.

Even with such savings, cloud could still prove too expensive, depending on the volume, complexity and standards of the data held by the provider, but, crucially, these factors stem from providers’ individual circumstance. Moreover, the shift to cloud for imaging can also require substantial investment in network infrastructure (e.g. local bandwidth) to leverage the benefit of cloud-based performance.  While there will be providers for whom AWS’ HealthLake Imaging product is still too expensive, the advertised and expected cost savings, will likely be enough to convince some providers, particularly when other factors, such as cybersecurity or the requirement to deliver capability across complex outpatient networks, for example, are considered.

Choosing Between Sellers

At present, the key differentiators between cloud providers are still minimal. While different providers may have different strengths, individual niches where they excel and particular partnerships that will ease certain use cases, any of the major cloud providers can, in essence, offer almost the same broad capability in cloud services for imaging. However, despite this comparability, leading cloud vendors are still beginning to better arm themselves and shape their identities in an attempt to build links to certain customer bases. Amazon’s focus on efficiency and the cost savings it offers is one such strategy, a play that, as highlighted, stands to place cloud capability firmly in the reach of acute and outpatient providers.

Other cloud providers also have their own strengths, however. Microsoft’s Azure finds itself in a particularly strong position, largely thanks to its acquisition of Nuance. Most obviously, that acquisition gives Microsoft a direct line to a claimed 77% of hospitals in the United States. However, that acquisition also fits in with Microsoft’s broader portfolio. It is, after all, not difficult to see the possible synergies with Nuance’s Powerscribe solution (and nascent, yet impressive DAX ambient reporting), combined with Microsoft’s ubiquitous tools, including Teams videoconferencing. This could bring ambient listening to all consultations and telehealth visits, leaving essentially every interaction structured and stored on the cloud along with relevant medical images.

Google, meanwhile, may lack the Nuance play that Microsoft can lay claim to, and it may lack the relentless operational focus that Amazon has developed through its commerce heritage, but its expertise in search, AI and broader image analysis, will give its own strengths, making it, for example, an attractive provider for leading academics focused on using their data libraries to develop their own AI algorithms.

Expected Arrival

In most cases though, these are concerns for the future. At present many providers aren’t considering long-term population health-focused imaging data repositories in the cloud, or developing their own AI tools. Instead, most providers are looking to the cloud for improved accessibility, efficiency, security and cost.

With these basics amply covered by all leading cloud providers, at present, which cloud provider hospitals choose is likely to depend more on customer-context, rather than unique capabilities. It doesn’t necessarily matter, for example, if a radiology department harbours a desire for an AWS imaging IT platform deployment if it is part of a large hospital network, which has just agreed an enterprise-wide deal with Azure. Almost all leading Imaging IT software vendors have some degree of flexibility on cloud-provider for hosting their applications, making cloud adoption often an enterprise, as opposed to departmental, decision.

By a similar token, hospitals in regions where there are restrictions on public cloud provision, where there is a preferred partner or a requirement for in-region datacentres, for example, have needs that trump any smaller local preference for individual cloud providers.

Despite these considerations, there is one area where AWS might have an advantage. AWS has arguably worked its way into a broader group of informatics partners (and larger market share) as “preferred” cloud provider, than some of its chief competitors. While some providers will disregard the partnerships their IT vendors have fostered, for many, simply adopting their imaging IT vendor’s preferred cloud provider partner will prove to be the most straightforward route to transition to the cloud, and as such, all else being equal, will be the one that is chosen.

There are some factors that will become increasingly important over time, such as the ability to manage and retrieve unstructured data, the ability to offer analytics so providers can use their cloud resource most efficiently, and even the adoption and ingestion of different data standards from across an enterprise imaging platform. However, in the near term, such subtleties are far from a provider’s priority.

In the near-term one of the main priorities, particularly for many mainstream providers, is cost. As such, Amazon’s claims of cost savings along with its repackaged and repositioned offering may make it an obvious choice for some. And for now, when fresh, first-time opportunities abound, that should be enough for the Seattle-based tech giant to deliver.

About Signify Premium Insights

This Insight is part of your subscription to Signify Premium Insights – Medical Imaging. The content is only available to companies that have subscribed to this paid-for service. To view other recent Premium Insights that are part of the service please click here

Signify Premium Insight: Google Searches for Imaging Success

This Insight is part of your subscription to Signify Premium Insights – Medical ImagingThis content is only available to individuals with an active account for this paid-for service and is the copyright of Signify Research. Content cannot be shared or distributed to non-subscribers or other third parties without express written consent from Signify ResearchTo view other recent Premium Insights that are part of the service please click here.

In its most high-profile bid to capitalise on the medical imaging market to date, Google has launched a new cloud-based medical imaging suite.

The suite is comprised of several tools, centred around Google Cloud’s image storage and Healthcare API suite, including NVIDIA’s MONAI for AI annotation and automation, BigQuery and Looker to help providers better navigate imaging datasets and Vertex AI to help accelerate the development of scalable AI models. Google says that by offering these tools in one product, it hopes to make diagnostic data more accessible and interoperable, while also readying providers for the development and implementation of artificial intelligence programmes.

While there has been significant promotion of the new suite, many of the tools were already available individually. What, therefore, does the launch really have in store?

The Signify View

Some well-justified scepticism will no doubt surround Google’s launch of its Medical Imaging Suite. The company has looked to expand its role in healthcare beyond feeding the imaginations of hypochondriacs several times, introducing Google Health in 2008 and discontinuing it in 2012, before rebooting it in 2018 and dismantling it again in 2021. It has, in the past also looked to develop front-line AI tools from its DeepMind division, consumer health products via its FitBit acquisition and apps for medical research among others. None of these tools, however, has yet made the significant, lasting impact that was expected by one of the world’s best-known firms.

Despite the sometimes inconsistency of these attempts, however, the company has been making headway in medical imaging with another, broader part of the business.

The adoption of cloud capability in medical imaging is still nascent, but some cloud providers, including Google, have become trusted partners for many earlier adopters. While Amazon Web Services (AWS) and Microsoft’s Azure are the public cloud providers that have seen the most uptake in medical imaging so far, Google Cloud, having reached some significant agreements with notable imaging IT vendors such as Visage, and Change Healthcare, as well as some notable and well-respected providers such as Mayo clinic, is hot on their heels.

Despite this, however, the Alphabet subsidiary’s presence in the market has been less visible compared to that of Azure and AWS. The launch of the Google Cloud Medical Imaging Suite highlights an end to this quiet and heralds the start of a more aggressive approach.

Remaining Relevant

Making such a transition has become increasingly important for the cloud provider. While Google has made progress, as AWS and Microsoft begin to pull away, Google misses its window to capitalise on the first phase of medical imaging cloud adoption. This is particularly true as Microsoft begins to capitalise on its Nuance Communication acquisition, for example, while Amazon continues to leverage its already extensive list of AWS partners.

Google will not countenance these advantages overnight, particularly given that on the face of it, its Medical Imaging Suite, which no doubt will be preferred by some customers with some specific use cases, is not a revolutionary leap. It may offer some advantages, but there is nothing truly ground-breaking that stands as a major differentiator compared to AWS or Microsoft.

That isn’t to say that there aren’t any aspects that aren’t attractive. The emphasis Google has placed on its AI offerings, for example, could swing some providers in its favour if they are looking to capitalise on their medical imaging data and facilitate its use among AI developers or indeed develop their own tools in-house. Its reputation for AI development could, in some cases aid its cause. This is especially true as many of the customers which have chosen to use Google public cloud are highly influential academic hospitals.

Reputation Management

Reputations work both ways though. While Google holds a staunch reputation for technical prowess, there are other factors that may give potential customers pause for thought. Chief among them are several high-profile incidents and agreements surrounding Google and identifiable patient data, including data from the Royal Free NHS Trust in London, and a deal with US healthcare provider Ascension. In these and other cases, Google’s actual culpability is somewhat moot, with the shadow of data insecurity, even if entirely unjustified, potentially enough to push a would-be customer in the direction of one of Google’s competitors.

Another concern for any potential customers considering turning to Google for their cloud provision is the vendors’ long-term commitment to medical imaging. While the more general aspects of Google’s cloud offering will continue to be supported, Google’s repeated high-profile salvos into healthcare, and the associated withdrawals, give the impression of a vendor that has no compunctions about pulling out of a market, reorganising its business units and ending its involvement in certain segments with little notice. Such an assessment may be unfair, particularly given that other cloud providers including Microsoft and IBM have both made equally high-profile pushes and retreats from some healthcare markets, but, with cloud representing a long-term investment, such concerns may weigh on decision makers’ minds when it comes time to signing on the dotted line.

These spectres are not impossible to exorcise, however. Google along with its peers are increasingly forging partnerships with imaging IT vendors in order to effectively create a joint sales strategy. Cloud providers, alongside vendor partners, are combining their efforts to sell to hospital networks, enabling the partners to highlight the benefits associated with a public cloud deployment, while also utilising the expertise from the imaging IT vendor in radiology.

Broader Responsibilities

Such evolution in sales strategy is also being mirrored in service provision. Along with the broader medical imaging market, deals are increasingly transitioning to managed service agreements. In terms of cloud deployments this is beginning to manifest as public cloud providers managing deployments much more closely, with for example, infrastructure and costing falling under the cloud provider’s remit.

Whether any of these factors are enough to sway a decision towards or away from Google, and indeed what influence they ultimately have on a providers’ choice of public cloud vendor, is still overarchingly dependent on individual deal context.

Google’s new Medical Imaging Suite will make the firm’s solution more attractive to many vendors, but any advantages will likely be overshadowed by much more significant influences. A deal’s locality, for example, may be a far more important factor in a provider’s decision if that provider is in a country which stipulates that cloud providers must have datacentres within-region, for instance, or if it is in a market sector that already has a preferred supplier.

As such, there are in most cases considerations far more significant than the differences between comparable cloud competitors. That, however, does not mean that Google’s latest efforts do not represent a significant step.

While the launch of its Medical Imaging Suite is unlikely to reverse the lead that AWS and Microsoft’s Azure have for public cloud departments, it does show Google’s intention. It highlight’s the vendor’s ambition in the space and lays the foundation upon which it can build over the coming years. Moreover, the launch also enables Google to remain competitive as other cloud providers such as Oracle and IBM which have already made their intentions clear, begin to more aggressively promote their own solutions.

Or, to put it another way, Google’s launch ensures it remains on the first page of search results, but, it has not yet offered anything to warrant rapidly climbing through the rankings.

About Signify Premium Insights

This Insight is part of your subscription to Signify Premium Insights – Medical Imaging. This content is only available to individuals with an active account for this paid-for service and is the copyright of Signify Research. Content cannot be shared or distributed to non-subscribers or other third parties without express written consent from Signify ResearchTo view other recent Premium Insights that are part of the service please click here