Project Northstar – Technology Deep Dive Webinar Thoughts? – General Discussion

bomalley

(@bomalley)

Posts: 12

Active Member

Topic starter

I would like to hear some of the GE Community feedback on the presentation on the Technology Deep Dive of Project Northstar yesterday. What are your thoughts on this new project so far and what did you think of the Orders demo that they did? I am honestly very concerned about this and that orders demo only made me more nervous with the slowness and refreshing. I like the concept that they are showing but I thought they would be farther along in the development process being that they are targeting a Q1 release date for this module. Also on interfaces, we have many order interfaces and I would like to know how those are going to change so that we can start preparing now.

Please let me know your thoughts!

Posted : July 14, 2016 3:29 am

lkeith

(@lkeith)

Posts: 457

Honorable Member

I would have to agree with you 100%. The slowness of the demo and the fact that you have to 'refresh' your screen to see the information is concerning. If it is that slow on a demo, imagine what it could be once their client base is all hitting the 'cloud'.

I am also excited about the concept of the rules engine and having the ability to now put in the rules for ABN checking as well as AOE questions but we have been burned in the past so not sure if this is going to be all that they say it is...i.e. the Advanced Specificity Problem Search that doesn't always work correctly so you have to have a work around for that.

Posted : July 14, 2016 3:58 am

mikeseale

(@mikeseale)

Posts: 344

Reputable Member

Honestly, if they don't have our data stored locally with some type of sync to the cloud, we most likely won't be able to use this. We don't get blazing fast speed to anything outside our network due to the geo locations of some of our clinics. The system will have to work flawlessly without the clinical decision support when we lose connection to the Microsoft/GE servers.

I couldn't tell where you could add orders on the new orders module, and if the form has to be refreshed manually, I don't think our providers will do it. Hopefully, it only needed to be refreshed on rule changes that are related to the clinical decision support system. Also, why couldn't he press F5 to refresh it like the HTML forms are able to do now? I hope we don't lose the ability to refresh with F5 for HTML forms in general.

I like that they are moving away from MEL although it concerns me that they haven't picked a programming language yet.

I like that they are going to have an open API, but I want to know the limitations of it. I would like to see if some of our existing interfaces that use HL7 to communicate can instead use the open API.

Posted : July 14, 2016 4:00 am

mond08

(@mond08)

Posts: 74

Estimable Member

I agree with all of the comments so far.

It seemed to me that the presenter was talking before hitting the refresh button almost as if it needed time before it would produce the results. Having to refresh would be a big negative.

I also agree with the fact that not knowing what will replace MEL is slightly concerning.
Does anyone know how this could work with/not work with Blackbird?

Dave

Posted : July 14, 2016 4:04 am

tdimitrov

(@tdimitrov)

Posts: 43

Eminent Member

Disclaimer: Venting/rant below 🙂

Underwhelming. Expected more from a "technical deep dive". Almost half of the presentation was high level stuff that we've heard before. Not much substance. The only thing that became evident was that the html orders module will appear as if it is native to the CPS client (even though it's served out of the HealthCloud) and will launch within the same window. The thing that I didn't care for was that the doctor has to manually refresh instead of that being done on the fly when they hit "OK".

I have a feeling that they are making it up as they go... They keep emphasizing that the reason why they are doing this gradually as opposed to "it's ready now, please consider migrating" is out of concern for their client base. I think the real reason is that at this stage this is mostly pretty mockups with little to no code behind the scenes.

The general vision and direction by going to the cloud is right. My doubts are centered around GE's ability to deliver (and deliver on time and issue free).

What about how printing, faxing, and as you mentioned, HL7 interfaces, etc. will be routed and handled... Bandwidth requirements, what browsers will be supported, IE version/feature dependency...and that's just scratching the surface...

Posted : July 14, 2016 4:05 am

virtualHITman

(@virtualhitman)

Posts: 319

Reputable Member

Amen and Amen

Posted : July 14, 2016 4:28 am

[email protected]

(@tdavidsonppcp-org)

Posts: 15

Eminent Member

Agree. I also found it interesting (and troubling) that for existing problems they would run a batch rules process overnight for patients on the schedule for the next day. They claim they will look at all the new results, etc. We get a lot of interface documents first thing in the morning. What will happen with results received after the batch process? What about same day appointments?

It was also concerning to hear they would handle interfaces on a case by case basis to create those that are not already "standards based". That will be a lot of our third-party interfaces. What about logic we have already build into Mirth or another tool? How will that be migrated. Is GE going to require us to pay for each interface we have to migrate? Anyone remember how long it took for GE to contact all their customers to run the CQR freshen up process? Weeks, and that was for one process. How long is it going to take to rebuild everyone's interfaces? Hopefully I'm wrong or missed something. I would have thought a lot more of the technologies and processes would be nailed down at this point. Seems like so many things are "being evaluated".

Posted : July 14, 2016 4:56 am

swhite

(@sbrooks)

Posts: 84

Trusted Member

I like the idea of a rules engine. We recently looked at both Athena and eCW. They both have rules engines integrated into the EHR. We recently had an on-site demo of Northstar. Honestly, we're trying to decide if we're going to stay with GE.
At our GE demo, the rep showed the rules engine working in the orders module to highlight missed tests. The module appeared to refresh quickly, and I would swear the orders module looked different than what I saw yesterday.

Posted : July 14, 2016 5:06 am

virtualHITman

(@virtualhitman)

Posts: 319

Reputable Member

Briana,

thanks for posting your question. I have a teleconference scheduled with Northstar's project manager this coming Monday (June 18) and will be compiling a list of concerns. I am personally not against moving to the cloud. What I am concerned about is GE's execution and migration strategy because as others have noted, there is a lack of detail in their presentations (not to be unexpected at this stage in the game). I am still giving GE the benefit of the doubt but the absence of clarity creates fear, uncertainty, and doubt for all of the end-users. My hope is that they would provide a forum for soliciting feedback and concerns from the CPS community at large rather than from a select group of large medical practices that they have hand-picked. I can update this group as things progress (assuming they do progress beyond the initial teleconference), but some of the general concerns and questions I have pulled together so far include:

No Offline Mode: It is imperative to have an offline mode if and when a practice loses connectivity to the "mother ship." It is unacceptable to only provide functionality when connected to the cloud. GE has not provided enough clarity on this issue. Side note: when we migrated from another EMR to Centricity, our selection committee eliminated every vendor that was entirely "cloud-based" because of connectivity concerns (and we have reliable Internet). You simply cannot bring medical practices to a screeching halt every time they lose connectivity or have latency-related issues with the Internet. GE admitted that the orders module would not work effectively without Internet connectivity. It is unrealistic and naïve to think that you can deliver a more reliable solution in the cloud, so you better have an answer for accommodating an offline mode.
Handling of Interfaces: Interoperability with local systems (e.g. lab information systems, RIS/PACS, and other medical devices) can be significantly impacted by cloud-based business logic. Latency is a common issue. As an example, our providers have grown accustomed to a very fast and reliable bi-directional interface for lab orders and results between CPS and our LIS. While the need for medical necessity checking and other rules checking at the point of order entry would be nice, it is frankly a distant priority for most providers especially if it is going to cause the interface to perform slower. I'm not saying that providers don't want this information (i.e. get a notification when they are about to order something that's not going to get paid); they just don't want their workflow to be encumbered. The demonstration of the rules engine was unimpressive and leads me to believe that the development team hasn't spent enough time in a real-world medical environment. There is no technical reason GE couldn't store the business logic locally and simply update it from the cloud. I am very interested in learning what GE will be doing to ensure the functionality and reliability of local interfaces are not compromised by their move to the cloud.
Migration Strategy - No Content Development Tools: It appears GE plans to migrate clients to the Northstar platform in phases starting with the CPOE module. It also appears that customers may have limited control over when this occurs. Even though a phased approach is necessary, no one likes living in a hybrid environment (HTML content and traditional data entry forms). I anticipate many users complaining about the inconsistency of the user interface as we go through this transition. What value is going to be provided that will help offset the pain of change? A more modern looking GUI is desperately needed but the value proposition is not compelling enough at this point. Many CPS users have developed highly sophisticated forms that deliver a high degree of functionality for their practices. The new orders module may look a whole lot better in appearance, but it will likely not accommodate the custom optimizations these practices have grown to appreciate. The transition to the Northstar paradigm would go much smoother if tools were provided to help folks develop content/data capture forms in the new HTML format out of the gate. Unfortunately, it appears GE has no plans for content development tools until much later in the process. I think that is a huge mistake. Most practices will see this as a step backwards rather than forwards. Appearance never trumps functionality.
Migration Strategy - Pricing: It is clear that GE will have to modify their entire pricing structure. My guess is that practices will pay for a base level of functionality to run CPS in the cloud; probably a monthly fee per provider. However, I'm predicting that to reasonably accommodate MACRA and patient/provider expectations, a practice will also have to add on "value added" services that will be charged by GE directly, or a combination of GE and other third-party providers. Tons of questions remain unanswered here. What will the pricing structure look like? How will the transition take place and when will it happen? How will this impact our relationship with our VAR? How will the costs compare to what we already pay in maintenance fees? I can understand that pricing models are complex and will take time to work out. I'm also sure that there are many other architectural decisions to be made before GE can begin to address the pricing specifics, but GE needs to understand that these questions will be in the forefront of every physician's and practice administrator's minds given the fact that most are working harder to bring in less revenue. Customers need a reasonable level of confidence that their cost of ownership will be comparable and scalable based on the needs of the practice.
Migration Strategy - Data Portability and Ownership: GE is after the data because they want to take advantage of "population health opportunities." They also need it to improve upon their "rules engines" over time. Conceptually speaking I have no problem with this and certainly see how everyone can benefit. GE has stated that even though your data will be stored in the cloud, the practice will still retain ownership; but as President Reagan once said, "trust but verify." My bigger concern is portability. Population health statistics and analytics are only as good as the data you are capturing. It is guaranteed that every practice is not using the same OBS terms to capture the same data for the same decision support purposes. In fact you can probably find practices that have used the same OBS term for different purposes over the history of their data capture, or have stored the same data in different OBS terms. That said, it is going to require some significant data scrubbing and transformation for practices to receive any value on the analytics side of things. My hope is that GE will provide some user friendly tools for practices to develop their own analytics. We found that analytics from CQR (GE's current analytic reporting tool) could not be trusted unless you followed GE's prescribed workflows to the letter, but that's not going to happen in a practice that wants to optimize or tailor the software to accommodate their specialized needs. The value of the analytics is contingent upon capturing good data and presenting it in a way that is actionable. I don't believe there is any way that GE can ever deliver to the expectations of their customers in that regard (at least not reliably), so customers need to realistically set their expectations. They simply don't have the bandwidth or the resources to translate everyone's data into a "common nomenclature." Maybe there are some common data elements around demographics, problems (ICD-10 codes), and PROCEDURES (CPT codes) that could generate some helpful metrics, but when you start talking about clinical quality measures (CQMs) that depend upon specific observations like labs or other OBS terms, you will have to provide tools to the customer so they can translate their data properly. In that regard, I do not envy GE or any other EMR company. On a lesser but still important note, I'm curious how document management will be handled. Where will all of the scanned documents and electronic faxes be stored? Will they still reside locally or will they also be migrated to the cloud? Will the customer have a choice?

There are certainly more concerns (security for instance) but the ones above constitute my top 5. I would love to see this thread expand. Do you agree with any of my points? What are some other concerns I should bring up to GE? Thanks in advance for any input.

Cheers,

GK

Posted : July 14, 2016 7:41 am

tdimitrov

(@tdimitrov)

Posts: 43

Eminent Member

Wow...impressive and comprehensive list.

1. Offline mode.

Absolutely valid concern. One possibility is to have circuit redundancy of some sort, ideally diverse physical paths. The secondary connection doesn't have to be of the same bandwidth as the primary one to save on cost, with the understanding and buy in from the leadership that things may slow down a bit until the primary comes back online. This may be a challenge in rural areas where connectivity is a problem and physical path diversity is rare. A distant third option would be satellite as a secondary/tertiary redundancy. I am not sure if other EHR vendors in the ambulatory space have such offline mode but I've seen it with EHR vendors in the hospital space - Meditech, Epic and others have solutions to accommodate circuit and power outage scenarios.

2. Interfaces.

I've been through an EHR move (hospital) from on premises to cloud where the interface engine stayed on the premises. This allowed quick communication with medical equipment and other entities. VPN tunnels were also terminated at a firewall on site rather than in the cloud. Not saying that this is the best way to approach this. The interface engine can certainly be moved to the cloud as well but that design decision also has other implications. We have no information how GE intends to handle interfaces, who will manage the interfaces, where the interface engine will reside, how will all these vpn tunnels/connectivity be handled, who would own it etc...

On the same wavelength as far as unknown variables such as pricing, scanning, and the rest of the issues you brought up... However, at this point I am not willing to give GE the benefit of the doubt. It is on them to convince me that there is hope at the end of the tunnel and we should commit to enduring the 3 year move to the cloud...which will more than likely take 4 at least, and if CQR is any indication, will be a painful experience.

Posted : July 14, 2016 8:53 am

bomalley

(@bomalley)

Posts: 12

Active Member

Topic starter

Greg,

Thank you so much for your post. This is exactly what I wanted to get started. I agree with all of your questions and concerns and I think GE needs to be transparent with us and give us these answers if they have them. I am not convinced that they know the answers to many of these questions yet and that is concerning. We also developed our own list of questions and sent to GE, some may be very similar to yours. We are hoping to have a call with them soon to discuss these as well.

1. We have many custom and third-party solutions integrated with CPS that requires the creation of custom stored procedures, views, directly editing data and extraction of data. Will all of these functions remain possible with northstar and will we be able to utilize our current database monitoring management tools including SQL Server Management Studio?

2. Will we be able to do performance tuning on our database?

3. Will we have our own instance of SQL? How will GE ensure poor performance and security vulnerabilities in one database does not impact our performance and PHI?

4. We take frequent transaction log backups and utilize lob shipping to replicate data to a DR site, test and reporting servers. We store the transaction log backups for 1 year and keep a monthly backup permanently. There have been a number of times we restored our test system to a point in time just prior to a crash in CPS to recover vital clinical data from a document damaged in the crash. Will we still have this level of granularity with northstar? How will disaster recovery be performed?

5. Utilizing a test copy of the CPS database we test new versions of the application for performance, workflow changes and new features. On most upgrades our physicians and staff will login to the test system to evaluate those workflows prior to implementing them into production. Will we have the same capabilities with northstar?

6. Will we have complete control on the scheduling of when upgrades to the online services take place? When we perform an upgrade we bring our DR site online to enable doctors access to clinical data in the event of an emergency. Although changes are not saved during the upgrade, the physicians always have access to clinical data (24 x 7). Will a emergency copy of the data be available during the upgrade?

7. We have many interfaces (through VPN connections) with external sources of documents (both in bound and out bound) utilizing QIE and the DTS server and other third party interface engines. Will we be able to continue to utilize these interfaces? Currently there is no cost to create an interface in DTS (we perform the work ourselves). In northstar will we be required to contract with GE every time we have interface needs rather than being able do the work on our own? If QIE (or other product) will provide the interface channel will be required to purchase a new channel for each interface?

8. Will externally attached documents to a patient chart remain in our facility or will they need to be pushed to the northstar cloud as well?

9. If the orders module is going to be released in Q1 2017 we need more information on how the orders interfaces will be changed in this cloud version. We need to contact our vendors now and begin the discussion of the interfaces changing. How will this new orders module affect billing going back into the PM system.

10. We realize that MEL is going away but will there be a need for form developing in northstar or will the form development in northstar be more rule creating that does not require a coding/programming background to complete?

11. Will we be able to continue to use Crystal reports via an ODBC connection in order to pull data into our own custom reports.

12. We have several forms that place or change orders based on criteria selected/entered into the encounter form. Will our current forms interact with orders without modification when the new module is released?

My hope is that GE will look at this thread and possibly see something that they have not thought about yet. Like you mentioned we all have been doing things the same but different. (IE: using different observation terms for different reasons. I know we have requested hundreds of custom observation terms for some of our forms, before they had observation terms for self management goals we requested some) How will that data get captured/migrated and integrated into their rules engine?

Thanks again for everyone's input! This is great.

Posted : July 15, 2016 5:01 am

mvanalmsick

(@mvanalmsick)

Posts: 13

Active Member

The demo was the exact same demo given at Centricity Live in May. I was hoping they would have more to show by now.

Posted : July 15, 2016 8:07 am

virtualHITman

(@virtualhitman)

Posts: 319

Reputable Member

Briana,

we had a brief teleconference with the project manager of Northstar (Serene Monroe) yesterday afternoon and I wanted to report back to this thread with a summary of the conversation. The purpose of the teleconference was really more of an introduction to her and her team and how we might get involved in the Northstar Project. Here were the key points:

She clarified that "Northstar" was the project name and not the name of a new product. The project is intended to move CPS functionality into the cloud (modularly over 3-4 years) and provide a higher level of interoperability between practices and other third parties (e.g. offsite labs, radiology centers, hospitals, specialized registries, etc.) instead of what is available today. I communicated the ongoing need for practices to be able to extend the product
GE expressed the need for "real world" customers to engage with them at various levels in the process (focus groups, beta testing, etc.) and are trying to figure out how to best go about doing that. She mentioned doing focus groups at CHUG and other regional forums throughout the country. This would give her team the opportunity to hear from customers about real interoperability challenges and some of the creative things they have done to solve or work around those challenges.
They asked for some initial feedback on the CPOE module and I expressed many of the concerns from my post below; particularly the high reliance on connectivity if business logic was only going to be maintained in the cloud. I also expressed my concern about interfaces with other systems (e.g. Lab Information Systems and instrumentation) and how the move to the cloud might impact that functionality. I expressed that performance and reliability couldn't be compromised with this transition or many physicians would see the Northstar initiative as a step backwards rather than forwards.
They asked about my prior experiences in providing feedback to GE and how they might improve. I told them how we had participated in the CPOE workgroup several months ago but that follow-up was lacking. That group had weekly teleconferences/webcasts over the course of about 6 weeks earlier this year and were told GE needed some time to deliberate over the data that was gathered and then they would get back to us. We never heard back from anyone so we had to assume the input from those involved was rolled up into the Northstar initiative. GE confirmed that that was the case.

Next Steps

GE expressed that they would like to follow up with us on two fronts:

They are going to have their technical team provide a "deeper dive" into the functionality of the CPOE module that will be offered via the cloud and solicit our feedback on some of the concerns and workflow issues around orders (being an endocrinology specialty, our practice has some challenging requirements around lab orders and imaging studies)
They are going to follow-up with me on my personal experiences in moving healthcare entities into "the cloud" as I have had a fair amount of experience in doing that in a former career.

I'll continue to provide feedback to this thread as it becomes available, but I also intend on talking to GE about placing a specific forum on the CHUG site exclusively for feedback on the Northstar project. I recommend GE assign several of their team members as moderators. I think it would provide a good vehicle for getting feedback from the multitude of practices out there and show GE how absolutely critical it is to not remove a customer's ability to tailor and extend the functionality of the software without undermining GE's need to control the back-end system. Our success is contingent upon their success and there are several practices and individuals (such as yourself) that want to be a part of the solution.

Posted : July 19, 2016 3:41 am

rwilliams

(@rwilliams)

Posts: 454

Honorable Member

I think the more likely concern would be an outage in the cloud taking every practice they have out at once. CQR is in the cloud and I get emails nearly weekly about QIE not being able to reach their servers. Cloud based seems nice for smaller practices but the larger you get the more you would want to control your own cloud instead of being lumped in with hundreds or thousands of other users utilizing the same servers as you.

What about performance issues?

Posted : July 19, 2016 11:52 am

tnc

(@tnc)

Posts: 517

Honorable Member

I have worked with some fine people at GE over the years. Some of these people are working on project Northstar and I am glad to see them there. The one thing which I have reservations about is support. The front line who is there for us, the customer, for assistance. I have worked GE since late 2001 and support has always been the wild card. Remember, a chain is only as good as its weakest link. We are now being asked, or told, to trust GE and everything will be OK. The latest iteration of "GE support-repair" from GE has been going on for the better part of 2016, likely in preparation for the Northstar announcement. That should be a good thing, for customers, to have better support. I had some conversations with some people at GE, working on the support problems. They asked for my feedback and I gave it to them no expectations. One of the things I mentioned was to assign a DID number (public phone number) to directly connect to the clinic down line so that when customers call in they don't need to waste the 2-3 minutes of waiting while navigating through the IVR system to make sure the press 2, then 3, then wait has not changed. Their response? They said that they think customers prefer to have it set up the way it is! So who here would like to call 911 for a medical emergency to get to the dispatch center and listen to a recording and not be able to get to a ringing extension for even a minute? Besides, you can leave the old system in place and assign that DID to the after hours ring group simultaneously. Am I missing something here? Of my list of concerns I did not expect anything but I thought at least this would be a no brainer.

For the record I am not anti-cloud. I utilize some cloud services now. One cloud I constructed myself. It spans two sites, has had 99.8% or so up-time (several outages over leased fiber beyond my control keeps me below the coveted 99.999%) but I do have RDP and VMWare client access to the equipment as well as limited after hours access to 1/2 of the points of presence and 100 percent access to the other if I need it. I have done this since 2008. If I were shopping for a cloud provider and had a choice, I would not choose a company with inconsistent support. My job at TNC has been to provide a solid infrastructure and keep that up-time very high. I have done that over the years and built a solid infrastructure but I can't lease fiber with SLAs from Tallahassee to Dallas or wherever that GE cloud will be and I hope support won't be worse than it is now (when we need them).

Lastly, I see that GE has cited the ability to use our data to do various things for population health care. There is no reason in the world that we need to send our data to the cloud for them to get our data. It could be replicated to GE just as easily with our existing CPS product. I don't see the point of parking my data thousands of miles away from our facility when we need exclusive access to it during business hours. If we lose access to that data, even for 30 minutes we are unable to provide patient care and are losing money. This is all assuming that the GE cloud will be for for 99.999% of the time. I hope it is but I also remember Centricity Business, and the first iteration of being allowed to choose a virtual infrastructure. GE had a very overpriced HP server with simple RAID, limited storage, no HA and some sort of software SAN from Left Hand Networks. What a fiasco. I chose to do my own virtualization and very glad I did. I heard the horror stories from more than one customer.

My wishlist-

I would like to see the development support standards based browsers (eventually no client needs to be installed). Start with dropping IE as the required browser (or Edge). We should be able to use Firefox, Chrome, Safari, IE or Edge if we want. That would make some of us happy to be able to use linux, OSX, or Windows clients (tablets, desktops or PC) at our choosing.

Focus on support. If that means increasing staff then do it. Train them in customer service and in the product.

Set up a customer knowledge base which we can use. If it is a known issue we should be able to go to the support site and find out so we don't have to waste time calling in, waiting, and maybe getting a call back from someone helpful.

communication-If there will be an update, we need time to preview/test/train our doctors and other users.

Mike Zavolas
Tallahassee Neurological Clinic

Posted : July 20, 2016 2:54 am

Welcome to the NEW CHUG— your ultimate destination for cutting-edge insights and discussions at the intersection of healthcare and technology.

Project Northstar - Technology Deep Dive Webinar Thoughts?