About the Authors: Matthew Holt is Co-Chair of Health 2.0 conferences (now owned by HIMSS), publisher of The Health Care Blog, where this piece originally appeared last month. He also serves as co-chair of Catalyst @ Health 2.0 and president of SMACK.health. Olivia Dunn, an associate at Healthy Communities Institute, and Kim Krueger, a Venture Associate at Plug and Play, also contributed to this piece. We are grateful to the California Health Care Foundation for supporting this work.
I’m happy to release an update to some unique data about a pressing problem–the ability of small health tech vendors to access data from the major EMR vendors and integrate their applications into those EMRs. For those of you following along, in 2016 when Health 2.0 first ran this EMR API survey, we confirmed the notion that it’s hard for small health technology companies to integrate with the EMR vendors. Since then the two biggest vendors, Epic and Cerner, have been much more aggressive about supporting third party vendors, with both creating app stores/partnership programs and embracing FHIR and SMART on FHIR.
In 2018, we conducted a follow-up survey to see if these same issues persisted and how much progress has been made. In this report, we break down the results of the 2018 survey and compare them to the results of our 2016 survey. As in 2016, survey response rates weren’t great, but in this year’s survey we asked a lot more questions regarding app store programs, specific resources accessed, troubling contract terms and much more. And if you look at the accompanying slides, we also pulled some juicy quotes.
The key message: In 2016 we said this, The complaint is true: it’s hard for smaller health tech companies to integrate their solutions with big EMR vendors. Most EMR vendors don’t make it easy. But it’s a false picture to say that it’s all the EMR vendors’ fault, and it’s also true that there is great variety not only between the major EMR vendors but also in the experience of different smaller tech companies dealing with the same EMR vendor.
In 2018, things are better but not yet good. A combination of government prodding (partly from ONC implementing the 21st Century Cures Act, partly in the continued growth of pay for value programs from CMS), fear of Apple/Google/Amazon, genuine internal sentiment changes at least at one vendor (Cerner), and maturity in dealing with smaller applications vendors from three others (Allscripts, Athenahealth, Epic), and the growth of third party integration vendors like Redox and Sansoro, is making it easier for application vendors to integrate with EMRs. But it’s not yet in any way simple. We are a long way from the all-singing, all-dancing, plug-in interoperability we hoped for back in the day. But the survey suggests that we are inching closer. Of course, “inching” may not be the pace some of us were hoping to move at.
All the data is in the embedded slide set below, with much more commentary below the fold.
It’s getting better but….EMR Vendors are still a bottleneck
Despite the new app stores and public statements, small health tech companies trying to integrate their applications with the big EMR vendors still regard them as difficult to work with, imposing unnecessary costs and putting up barriers to data access. But when we asked how they compared to 2 years before, 63 percent of small health tech companies thought that EMR vendors were making a modest improvement in allowing easier API-access and other forms of data access for third-party tool integration, while 23 percent thought the improvement was significant. When asked the same thing about providers (hospitals), only 54 percent said modest improvement, while 23 percent said providers had improved significantly. Worth noting that we didn’t get a response from Apple, which now has data access for over 100 hospitals! (Helps to be big and rich, and not just with hospitals! Cough, cough, FDA.)
Most respondents believe that providers want to get them easier access to data, but they are often hindered by complex bureaucracy or confusing technology. In general, it still seems that data access is a function of a provider’s willingness to push their EMR vendor to open up. From the small tech company point of view, EMR vendors seem to be under no pressure to integrate and few providers are putting pressure on their EMR vendors. To quote one respondent “Still a lot of talk, obvious value to be had, but limited actual progress.”
And to be fair, we’re not sure small tech companies are helping themselves too much. Since 2017, Health 2.0 and HIMSS have featured the Carin Alliance a bunch, and Aneesh Chopra, Ryan Howells and their gang have been banging the drum on small companies getting access to data, yet when we asked about their work, 80 percent of small tech company respondents said “Who is the Carin Alliance?”
But third party application integration is becoming a bigger deal (thanks, Redox!)
Whether or not sentiment is mixed, the data about what’s actually happening is clear. Many more small tech companies had actually completed integrations with big EMR vendors, although the complexity of those integrations were perhaps less than what was being done before 2016 (most just read and write and fewer are trying to manipulate data within the EMR). And 85 percent of these integrations were started after early 2016 (when the previous survey was in the field).
As you’d expect given its market dominance, more vendors have integrated with Epic (64 percent), followed by Allscripts (54 percent) and Athenahealth (53 percent), with Cerner at 45 percent befitting its later start at getting the API religion. How they accessed the data shows that 3rd party integration vendors (Redox, Sansoro and a few more) are becoming important. Most got to Athenahealth and Allscripts data via their APIs, but using 3rd party integrators is becoming more important especially for Epic and even Allscripts.
But don’t think it’s easy. One respondent wrote, “For many vendors, it is a combination of ways to get all the workflows we need (other than Allscripts and Athenahealth). Epic still requires a mix of API, HL7, direct message and batches. Others are HL7 and direct.”
Now the good news: EMR vendors are becoming better partners
In 2016, most small companies thought that few of the major EMR vendors were generally supportive about their integration efforts. The two big exceptions then were Athenahealth and Allscripts. In 2018, most everyone is doing better, with Cerner making the biggest strides (61 percent now saying Cerner is supportive vs 31 percent in 2016)—that’s consistent with what we see from them both in the market and at our conference. eClinicalWorks, Epic, Meditech and GE are still below 50 percent but are all getting slightly higher marks.
Epic, though, is still its own animal. Eighty-three percent of the respondents say a large provider client is necessary to get them into a conversation about accessing data. Cerner wasn’t much better (69 percent), while for Athenahealth, Allscripts and McKesson (which is now mostly under ChangeHealthcare), a large provider client was not needed.
There was also a large minority (40 percent) who found the contract terms of the vendors, e.g around intellectual property, troubling.
When it comes to actual partnership programs, things are looking up. In terms of the small health tech companies finding a partner program that exists and is easy to use, Allscripts (63 percent vs 43 percent in 2018) and Athenahealth (72 percent vs 43 percent) are doing much better and even Epic has gone from 52 percent saying the partnership program was non-existent in 2016 to only 23 percent in 2018.
What about those App Stores? Programs are taking off despite high costs
Probably the biggest change since 2016 has been the creation of iOS and Android-like app stores by many vendors. Plenty of smaller companies are taking part. Fifty percent of respondents are in the Athenahealth MDP marketplace, 39 percent have taken part in Allscripts Developer Program and 33 percent are in the Epic App Orchard. But there’s lots of whining—especially about the cost. As one respondent wrote, “As an earlier stage start up, having to pay 20 percent revenue share for a client clicking through to us from a marketplace with virtually no sales support is high but necessary evil.”
Additionally, we heard a lot not only about the cost, but also about the terms for Epic’s App Orchard. This included Epic wanting to know the details of their technology, being unfriendly to partner vendors and not open to any negotiation of terms.
Since we took this survey (but before we published it), this message seems to have got through, with Epic promising last month to reduce the costs of their App Orchard program.
You got the partnerships and permissions set up? Now integrate!
When asked about specific vendors’ support for their actual integration efforts (i.e. the bit after the partnership puffery), things haven’t changed that much. In general Athenahealth (89 percent) and Allscripts (78 percent) are helpful, Cerner in the middle (52 pewrcent in 2018 vs 48 percent in 2016) and everyone else regarding as more hindering than helping.
But since our 2016 survey, API quality from the big guys has generally improved. Now we’re not getting too carried away, most still feel EMR vendors’ APIs are “not great but workable” and only EMR vendor looking really good was Athenahealth (85 percent saying “APIs were of high technical quality”).
Most of the time (in an increase from 2016) EMR vendors charged a fee for API access but these are mostly now settling on a percentage of revenue share and was less than $25,000. We think that indicates that the app store model is taking hold.
At Health 2.0 we’ve been trying to shine a light on this topic for some time, and the good folks at SMART on FHIR (the Mandel/Mandl twins, Zak Kohane et al) and the Argonaut/FHIR/CARIN crowd (Graeme Grieve, Aneesh Chopra and a cast of hundreds), have all been banging the drum as well as laying down great work for several years. And yet it’s health tech, so slow incremental progress is probably what we should expect. The state of play is that the big vendors are all now awake to the issue, but there’s lots to sort out before access to data and integration into APIs becomes as automatic as we see outside of health care. Patients with complex diseases still have multiple portals often into multiple version of Epic, and leading journalists are still writing stories about having to have tests redone because they can’t get the images or data to cross the street. But I get the sense that the levee is sprouting leaks. Cerner and Allscripts are moving most on-site installations to public and private cloud, AWS and Google Cloud are sniffing around as data storage providers and starting their own partnership programs. Epic remains Epic, but has an app store and is reacting to some of the criticism.
I’m also hoping that this type of a survey will soon become irrelevant because the topic about access to data and ability to integrate applications will soon become one of those things that we wake up one day and realize aren’t problems anymore. We’ll check back in a couple of years and see how close that day is.