Have we been successful or not?
As healthcare consultants and HIE experts, we get that question often from leaders and managers in healthcare about HIE and interoperability when discussing healthcare systems integration.
Let’s say we’ve certainly learned some things in the journey towards interoperability.
The Data Dilemma
We know, for example, that not all data is the same – and without data you have very little value. Which means you can spend too much on data, or too little, and end up in essentially the same place.
The need for data certainly caused many organizations early on in the interoperability journey to attempt to get all the data they could through spending millions of dollars on integration. Very few have been successful in getting all the data and now we have many HIE’s in jeopardy of dying because once the federal grant money runs out there is little value in only having small amounts of the data that doesn’t affect care quality or cost. This does not apply to all healthcare integration projects, but we’ve experienced it in many HIE’s in the market across the US today.
This situation was predictable as we’ve seen the pattern before. When a perceived innovation that may generate serious value is introduced, passionate activity can overrun strategic growth . When HIEs first entered the market, there were very few standards and none really beyond HL7 to utilize in healthcare integration systems and for aggregating data. Products like Novo Grid and Access Medical were popular solutions to deploy in getting data captured from systems through screen scraping and screen capture. Companies like Medicity, ICA, Axolotol, CareFX and dbMotion were dominant players in the big HIE space.
Where do we go now?
Defining Data for Effective Usage
We have some assets with existing state designated HIE’s, regional information organizations, and hospital or healthcare system based HIEs. The first step in resolving the data dilemma is to define the value of different types of data in specific circumstances.
For example, when does an EKG matter and how much does it matter?
There are times when the last 3 lab results trended and graphed in a presentable format served up to a physician or care giver with the patient in front of them is needed. With new standards in CCD and its’ derivatives we have a chance to exchange data more readily and answer these kinds of needs. But before we jump in and make exchange technology work more than it does now – we need to step back and strategically ask what data is really needed in what situation.
It isn’t as simple as saying, “if I have a patient in front of me who is having chest pain, and who had an EKG this morning along with lab tests, that I need that EKG and the test results or I will have to repeat them at a cost equal to new tests.” The healthcare systems integration strategy also needs to account for the time to diagnose, physician and staff time, and the impact on the patient of repeating procedures.
Conversely there are times when a chest pain patient won’t require these same pieces of data because they have a history of anxiety.
That seems like a complex problem to figure out, but what it really involves is understanding not just workflow patterns but data use patterns. Once we begin to understand how the data is either used or not used for any given situation, we can map it and create patterns that are usable in determining a priority set of data to be made available at the time and place of care, whether that is in-front of a doctor or directly to a patient at home. We can then make sure we provide only the necessary data to care providers so they can be more productive in diagnosis and treatment. The larger issue of sustainability of health information exchanges depends in-part on redefining them from standard data store and forward technologies to data and information management technologies that support basic patient/provider use-cases to analytics, to compliance interoperability, and clinical data interoperability.
Here’s an interesting need for interoperability that has recently arisen – in the compliance space. From HEDIS reporting to just moving census data between payer and provider there are a lot of use-cases in which interoperability is needed with data that overlaps the clinical and non-clinical spaces. For example, getting ADT data alerts are important in the clinical space to understand when patients are discharged and where they go. Likewise, it is also important in the authorization and compliance process in getting approvals to repatriate patients back inside a narrow network. There are many other examples where data would be used in both situations and there is a need to use interoperability strategies to solve these problems. The use of data within workflows that make sense is important and we must define it before we can effectively deliver on the promise of HIE.
Data Delivery from Epic Solutions
Another data challenge to overcome is that we need to actually have data before it becomes a solution or a problem. In the market today, we have one dominate EMR player who offers the ability to share and exchange data between disparate systems; Epic.
Epic’s Care Everywhere allows the exchange of information between Epic Systems as well as CCD exchange between Epic and non-Epic systems. In our capacity as consultants, we see challenges every day in which connectivity issues need resolution because a CCD interchange won’t work on one end or another, or custom work has to be done in the cases where CCD cannot be used due to vendor readiness. In those instances, custom interface work is often needed. This custom work can be as little as a 40 hour process costing $8000 or as much as $120,000 taking months.
The question all healthcare systems using Epic face at some point is when to utilize which particular solution. Sometimes urgency requires custom interface development regardless because it just has to be done. As an example, when patients need to be repatriated from one hospital to another due to narrow network requirements then you have to have real-time ADT data or costs rise while waiting on information telling you where your patients are.
What Are Your REAL Data Needs?
At Orchestrate Healthcare, we recommend that you really look at your data needs more closely now then ever when planning your healthcare systems integration. Some doctors will tell you they want all the data. That requirement is actually easier to meet then most, but “all the data” quickly becomes a workflow problem and they will want it undone when that point is reached.
Imagine this conversation:
“No, don’t give me ALL the data, just give me the bio-markers AND anything else critical.”
What else is critical doctor?
“It depends on where I am and what type of patient I have.”
We agree, and this is clearly the next step in our evolution as data aggregators who apply tools to data to derive decision support. If data aggregation, normalization, and the application of ontologies didn’t have the potential to really change healthcare we would probably just try to meet the Meaningful Use criteria, take the dollars, and continue on. But correctly identified data in the right place and time can provide critical answers. As a nation of healthcare users and suppliers we need both data and answers.
Given the problems we have already described with data acquisition and the progress to date the next real question is:
When will it change so that we can resolve the problem?
As strong as Epic is, it is unlikely we will ever have an all Epic setup so that data can be interchanged without some level of integration. The process we recommend starts with:
1. Defining what data really brings value, then
2. If you have a health information exchange system, either import CCD or its derivatives or work on direct interfaces.
To date, we have seen substantial problems with simply exchanging CCD’s or more practically sending CCD’s through unidirectional transfers. One of the issues with CCD has been the inability to parse them without custom work being done. In the absence of the ability to parse CCD’s you should pursue a strategy to get the data via HL7 interface until the technology evolves. You can also use manual strategies to parse the data in order to normalize it and make it useable in a trendable form. Getting data into the system in a useable form is far more important than the cost of doing so at this point. That is why it is so important to choose only the data you really need up front so as not to expand the cost unnecessarily.
Technology Advances Warrant Action
The technology is changing and we will see rapid change in our ability to operate bi-directionally over the next 5 years. The ONC’s 10 year plan is to get to that point but the technology change itself will get us there faster than that. Unfortunately the rapid change in healthcare itself is happening right now and affecting reimbursements as well as needed cost structures. Having the ability to do population health management, disease management, and more specific predictive analytics will be the key to getting those cost structures down to the level where hospitals and healthcare systems can avoid the destructive losses associated with major declines in reimbursement coupled with increased regulatory compliance. Sitting still is not an option right now because sitting still means falling behind against healthcare reform.
If your healthcare doesn’t already have a plan like we have suggested above for healthcare systems integration, then the recommendation would be to follow this process:
Value your data
1. Literally place a value on your data to understand what data is needed where and when and by whom in a more exact way than you have done before.
2. Analyze what benefits you get by presenting the data in those situations. Things like reduced redundancy, surgery delays, over treatment, under treatment, productivity impacts to physicians and other providers of having the data or not having the data, and how having the data affects the treatment process or diagnosis of patient issues.
3. Figure out how much data you can get of what you need from existing sources including the distribution method of getting the data where it needs to be and if it can be done in the time status that meets the need. As for the data you don’t have access to; that also needs to be resolved.
CCD transfer through interface engines, HL7 direct integration, and custom integration are all options that can be supplied by companies like Orchestrate Healthcare. Our KLAS ratings are evidence that we do it better than others, but what is even more important is the quality of the engineers you have working on the problem. There are so many challenges one faces when working on any of the three types of interfaces that talent really matters. You need someone who understands project management, the data itself, how to write and test and effective interface, and someone who understands the importance of maintenance in the process. Having in-depth knowledge of HIE’s themselves is also important because hopefully the interfaces you are building will end up feeding an HIE for normalization and storage to go along with the distribution of the data.
We Are Successful With Healthcare Systems Integration
If the definition of success in healthcare systems integration is that data is presented as information for use by providers in delivering higher quality care at a lower cost; then an organized and well executed process is necessary. Contact Orchestrate Healthcare today and we can help you with your strategic plan, estimate the work process and costs in delivering the interfaces, and then actually do the work.
We believe that a population health and disease management 2.X version will arrive shortly that will allow for much better risk prediction and patient management. But to realize cost reductions and quality improvement value you have to start with the data and provide it to these new tools on your way to a better, bigger, data value for your organization.