I think it’s quite easy to predict that 3-5 years from now, one of the top topics on this blog and in the EHR world as a whole is going to be around EHR data extraction or if you prefer EMR data conversion. I’ve previously predicted that by the end of the EHR stimulus money we’re be lucky to achieve 50% EHR adoption. So, you’d think that in 3-5 years we’d still be talking about EHR selection and implementation. Certainly, that will still be a topic of discussion. Not to mention, which EHR vendor they should go to for their second EHR. However, I am certain that 3-5 years from now we’re going to see a mass of doctors switching EHR vendors.

As part of my EHR blog week challenge (if you’re a blogger, you should participate too), today I’m going to highlight one of the foremost EHR professional and technical services company’s blog, Galen Healthcare Solutions which focuses on EHR data conversion.

I know I’ve written about EMR data conversion a number of times before. Although, I haven’t written about it much for quite a while. I guess meaningful use and the EHR incentive money has kind of dominated the conversation. However, there’s much that can and should be said about EHR data conversion.

The first thing anyone should know about EHR data conversion is that it’s not easy. In fact, it’s quite frankly an incredibly painful experience in almost every regard. Just take a look at this blog post summary of the EHR Clinical data conversion process by Justin Campbell of Galen Healthcare Solutions. He summarizes the steps as follows:
* Data Extraction
* Data Analysis: Cross-Referencing
* Design: Data Filtering, Matching (Provider, Patient Item), and Exceptions/Errors
* Testing
* Go-Live

I believe the most challenging item on this list is likely the Data Extraction. Sure, the data analysis and design are a pain to do and do well. However, the data extraction is often the most difficult part of an EHR data conversion, because you’re often working with an unfriendly EHR vendor that has lost you as a customer. Unfortunately, many EHR vendors haven’t heeded my call for EHR data independence, and so it can be a miserable experience trying to get the information and access you need to do an EHR data conversion. In some cases the EHR vendors will try and hold that data hostage.

The key for those selecting an EHR software is to be sure that the process for exporting your data from the EHR is part of your EHR contract. If it’s not, then add it to your contract. If they won’t add it to your contract, there are 300+ EHR vendors to choose from. Certainly it’s a part of the EHR contract that you hope to never have to use. Don’t take that risk.

Justin Campbell has also posted a few different data conversion success stories on the Galen Healthcare Solutions blog. Obviously, Galen has a lot of experience with the Allscripts Professional EHR software and so you’ll note this bias throughout the blog. However, the experience of the conversion is very interesting.

Here’s a paragraph from one of their data conversion success stories: Azalea Orthopedics.

To facilitate this conversion, flat-file extracts were obtained from MedManager for dictionaries, demographics and appointments. However, instead of using these extracts to import into Allscripts PM, an alternative approach was taken in which real-time appointment and demographic interfaces were deployed from the client’s existing Allscripts Enterprise EHR to the new Allscripts PM environment. This offered the flexibility of having the PM data populate real-time. Interfaces were also required from Allscripts PM to Allscripts Enterprise EHR. Thus as part of the go-live, existing reg/sched interfaces from MedManager to Allscripts Enterprise EHR needed to be deployed.

I have to admit that this kind of complexity in healthcare is what drives so many doctors nuts. I’m sure there were some functional reasons that they had to do all these interfaces between the systems. What I don’t understand is why the interfaces need to stay in place after the conversion is complete (at least if I understand it correctly). Did Galen really have to implement an interface between Allscripts PM and Allscripts Enterprise EHR? I’m sure there’s some long history for why this has to happen, but it’s such a terrible design. Certainly this isn’t Galen’s fault, but Allscripts. Interfaces are really great….when they work. When they don’t work, they drive a clinic, the IT person and even the EHR vendor absolutely nuts. I’ll be interested to learn more from Galen about why they did what they did.

I did find their report on the number of transactions processed fascinating:
Demographics: 156,900 processed in 491 minutes (8.18 hours)
Appointments; 313,280 processed in 1570 minutes (26.17 hours)

That’s a lot of data being processed. Can you imagine having to run the 26 hour data conversion twice if you messed it up the first time? Yep, data conversion is a tricky thing and can be very time consuming if you’re not really thorough in the process.

Imagine how much data will be collected 5 years from now with all these EHR implementations happening. Plus, the above data was only appointments and demographics. It doesn’t even include the physicians charting and other clinical data.

Two articles I’ve written in the last 24 hours have gotten me thinking that we’ve already entered the post-implementation era of EMRs, even as implementation remains in progress at so many healthcare organizations. While the vast majority of hospitals and physician practices in the U.S. still don’t have full-featured EMRs in place, many are already looking well into the future.

As you may already know, HIMSS on Tuesday released its first-ever survey on “clinical transformation.” According to HIMSS and survey sponsor McKesson, “Clinical transformation involves assessing and continually improving the way patient care is delivered at all levels in a care delivery organization. It occurs when an organization rejects existing practice patterns that deliver inefficient or less effective results and embraces a common goal of patient safety, clinical outcomes and quality care through process redesign and IT implementation. By effectively blending people, processes and technology, clinical transformation occurs across facilities, departments and clinical fields of expertise”

As I reported for InformationWeek, 86 percent of organizations surveyed had a plan for clinical transformation in place or at least under development, and just 12 percent of respondents called organizational commitment a barrier to reporting on quality measures. And though nearly 8o percent indicated that they still gather quality data by hand and 60 said they don’t capture data in discrete format, more than half already had software specifically for business intelligence. This tells me that analytics is here to stay.

I kind of knew that anyway, since the bulk of the program at last week’s Wisconsin Technology Network Digital Healthcare Conference was devoted to BI, data governance and advanced analytics tools, even in the context of Accountable Care Organizations. (My story about this for WTN News appeared this morning.)

“I’m ready to declare the era of business intelligence,” said Galen Metz, CIO and IS director for Madison-based Group Health Cooperative of South Central Wisconsin. Though he criticized the proposed ACO rules for being too “daunting” for the average provider, Galen and other speakers said that it’s time to harness all the new, granular data being generated by EMRs and, soon, ICD-10 coding.

It may seem “daunting” now in the midst of all the preparations for ICD-10 and meaningful use, but it’s good to know that many healthcare organizations see a light at the end of the tunnel and know that the future bring better healthcare information in exchange for all the hard work and investment today.

My last post on EMR companies holding practice data for ransom was very popular and had some very interesting discussion in the comments. Honestly, every EMR vendor should be considering the impact of the choice to not liberate the data in the EMR.

No, I’m not talking about being loose with the data. HIPAA will come back to bite you if you do that. Plus, no doctor will want to use your system. What I’m talking about is making the data in the EMR available to the doctor. In fact, if your a doctor or practice manager reading this post, you should make this a requirement of the EMR vendor you select. If you already have an EMR vendor, you should work to have them incorporate this feature in their EMR ASAP.

Lest you think that I’m just being pro doctor and not considering the EMR vendor, let me provide you some business reasons why an EMR vendor would want to create a plan to make the data in their EMR available to their doctor in a liberated format.

As one EMR vendor put it:

Bottom line, its a good business practice to provide the data in an accessible manner to the client when and if he/she wishes to move on to another EHR.

Let me add the following:

If you don’t hold the EMR data ransom then you will have a better image in the community.

1. If someone wants to leave and you make it hard, word will travel that you held them ransom and others will be afraid to choose you because they know they’ll be locked in. We all know how tight the medical community is.

2. If you free the data, users will trust you more cause they know they can leave at any time. Plus, they see you’re doing what’s best for the customer. This narrow minded focus on the customer’s needs will certainly carry over into other areas of your application and lead to happy customers who can leave, but won’t have any reason to leave.

3. SaaS EHR vendors in particular should offer this service. One of doctors biggest complaints about SaaS EHR is their fear that their EMR data isn’t stored in their office. What easier way to allay this fear than to provide them a regular copy of their data?

Do NOT underestimate the power of your image with the customer. It leads to happy long term customers, but also leads to future customers. I saw a recent study (which I can’t find right now), but it was amazing to see the amount of influence that other colleagues EMR recommendations made on a doctor’s EMR selection decision. Providing doctors their EMR data which may lose you a few customers along the way, but it will retain and find more customers than you lose.

JamesNT wrote a really interesting forum post recently about how a number of EMR vendors are holding doctor’s patient information “ransom” (his word) from them. Here’s his whole description and he even names a few EMR vendors and the challenges related to getting the EMR data out of their systems:

To many EMR’s lock up the practice’s data and hold it for ransom. The data entered into an EMR belongs to the practice, not the EMR. It is not fair for EMR’s to not provide ways to interface or export data from the database. If a doctor wants to hire an IT person or developer such as myself to write custom reports or export data from the EMR, then it should be possible. Consider the following examples:

Amazing Charts: They use SQL Server 2005 Express as their database but they remove the built-in Administrator account from the SQL instance and change the SQL Server SA password. This means anyone hoping to interface or export data is at a loss – and Amazing Charts will not share the SA password. Amazing Charts also does not publish a database diagram.

eClinicalWorks: Overly complicated database. Does not publish mySQL password (you can find it, though). Does not publish database schema. If you ask them for help, they want to charge $5000 to build an interface.

PODMED (now TrakNet): Kudos for sharing the SQL Server SA password – but does not offer a published database schema.

GE Centricity: Database schema available – if you are willing to tell a bold-faced lie to someone to get it.

Medinotes: Even after sunsetting the product, Allscripts refuses to give out the ODBC driver and database password.

MD Logic: Uses a pathetic HL7 file interface. You can place only one patient demographic in each file – so if you have 200 patients to update that means sending 200 files.

Officemate: Uses SQL Server and it is easy to get to their database – but they do not offer the schema.

I find this situation deplorable. Every EMR should make it easy to get to the data and not try to hide it or charge outrageous amounts for an interface. Seriously – who here would pay $5000 to make an interface?

Of course, he’s just highlighting the EMR software he’s used. I’m sure there are hundreds more EMR vendors like this.

Then, there’s also EMR vendors that don’t hold your EMR data for ransom like Medtuity. Here’s what Matt Chase from Medtuity said about what they provide to users of their EMR:

At Medtuity, we provide open access to the SQL database. We also provide an export facility under Options. You can export each and every encounter, years and years worth if you wish, to a PDF file for each visit, neatly labeled with the date of the encounter and pt’s name to keep it from colliding with other PDF documents. You can also export a CCR for each pt.

We also have our own proprietary format in XML. For a group with a huge number of records, they may wish to hire a consultant to write a program to consume that xml into a new system. Our xml format is most complete and includes the stuff you would not usually wish to transfer (the audit trail on that chart, for example). But it is there. We also have CSV format, but let’s face it, you cannot export sophisticated data in a CSV format. It’s fine for demographics.

How “liquid” is the data in your EMR software? This discussion is a very important one between you and your EMR vendor when you’re selecting an EMR. Make it part of your EMR contract.

More EMR vendors need to voluntarily step up to the plate and provide this type of EMR data liquidity.

Today, I came across the best description I’ve seen of the difficulty of storing healthcare information and also making that information shareable with another EMR in a way that is meaningful.

Longitudinal patient information is arguably one of the most temporally and spatially complex information sets known. Certainly GIS and others are complex as well but the science of medicine and therefore healthcare is constantly changing creating a moving context. To understand how to treat a patient the healthcare provider needs to be able to understand what has worked as well as what hasn’t worked in the context of what was known about the patient and the treatments available at any point in time. This creates an environment of very complex data relationships. If any one of those relationships are broken then the semantic context of the data is lost and now there is a loss of information. Data items need to be bundled and stored as a complete unit of understanding for them to constitute information. Once broken apart into separate data items they are much like Humpty Dumpty.

Certainly the above description describes the challenge of storing and sharing both the healthcare information and the context of that healthcare information. I still can’t help but think that we need to simplify our goals for EMR data sharing into small achievable goals.

The above description also kind of reminds me of my previous post about the “Body of Medical Knowledge Too Complex for the Human Mind.” The description above reinforces this challenge as well.