Statement of Matthew King, M.D., Chief Medical Officer, Clinica Adelante, Inc, Surprise, Arizona
Testimony Before the Subcommittee on Health
of the House Committee on Ways and Means
July 24, 2008
Background
Clinica Adelante, Inc (CAI) is a Community Health Center located in the Phoenix, Arizona area. We have seven sites that serve both urban and rural populations and a mobile clinic that serves remote areas of Maricopa County. The clinic has 26 providers, including family practice, pediatricians, internists, OB/Gyn, mid-levels, and dentists. We see about 32,000 individual patients annually and about 90,000 encounters. 50% of our patients are uninsured, 40% Medicaid, 3% Medicare and the rest commercial insurance. We provide sliding fee services to those at 200% FPL or below.
In 2000, I took over as Chief Medical Officer for the clinic. CAI was engaged in National Chronic Disease Collaboratives sponsored by HRSA. We used Wagoner's Chronic Disease Management Model[1] to improve care for some of our diabetics and asthmatics, which has been successful in showing dramatic improvements in chronic disease outcomes. The model utilizes patient education, nationally recognized treatment guidelines, a rapid process change model known as PDSA cycles and a chronic disease registry. The registry is a critical piece of the model because it can be used to track the population and also provide a means for outreach. However, it is not designed to be used in the exam room with the patient, so the patient data needs to be entered manually into the registry later. This double entry of data--once in the exam room and once in to the registry--is error prone, time consuming and costly.
Our desire was to extend the model to everyone that walked into the door so that each patient could have their own personal health plan based upon their age, sex, risk factors and disease states. However, we faced two main challenges. First, because the registries required double entry, we estimated that we would need to hire 24 more data entry specialists; however, we did not have the funds to do so. Second, the time required to do the preventive health would have a negative impact on our revenue. We knew that we needed to find an EHR solution that was relatively inexpensive and could support data entry into a registry without double entry; because it could be used at the point of care.
The Search for an EHR Solution
We started a search for an EHR. The search was disappointing: The products were very expensive, between $200,000 to $500,000, and they really didn't perform chronic disease management out of the box well without expensive customization; and they were deployed in a consumer unfriendly environment that included consumer hostile contracts, vendor lock, poor interoperability, and a licensing and support structure that negated the natural leverage of collaborative networks. Because of my prior exposure to Linux and other open source products, I wondered if there were open source solutions that would address the clinic's needs.
I would like to stop for a moment to discuss what Open Source means in the context of Health Information Technology (HIT). Open Source software allows one to see the source code and is freely available. The Open Source license used by organizations such as WorldVistA guarantees that not only is the code available to be examined, it is also available to be enhanced by the community and the enhancements cannot be lost or trapped in a proprietary product for the sole benefit of one vendor and its customers. Improvements must be donated back to the community of users. Enhancements to the code can come from volunteers, vendors, funded projects, IHS, VA, etc. These enhancements are checked by experts and only released after review. The important points here are that innovations can come from many sources, collaborative development compounds the value and effectiveness of investments, and the processes are transparent, organized and safe.
The following is a list of what we perceive through our direct experience to be some of the key benefits of the open source model in healthcare:
1) Software quality and standardization accelerated by transparency - The transparency of the code assures better software quality and conformance to coding standards and security. Security flaws are more likely to be found and quickly addressed, often within hours of discovery. Non-conformance with open standards is not tolerated by both developers and users.
2) Rapid innovation and improvement -The improvement cycle needed to keep the software current in response to the dynamically changing healthcare environment is much more rapid than in proprietary business models.
3) Improvement driven by user needs - Enhancements and fixes are directly driven by what users need, not by marketing, shareholder or other non-healthcare related priorities. Community Health Centers, for instance, can drive changes to update their UDS reporting, while a proprietary vendor might not have the business case to make the code changes.
4) Lower total cost of ownership - No licensing fees mean less upfront and lower total recurring costs.
5) Competition focused on service excellence - Flexible support fees mean greater chances to leverage technology. For instance, if support fees are fixed by number of servers, not providers, every provider assigned to that server will spread the costs over more and more users. In the traditional model, every provider added to the system will cost another license and more support fees.
6) Collaborative leveraging of resources to improve “products” - Open source means quality management tools, clinical tools, interfaces, training and deployment materials are all shared. Going forward, the costs to participate are less and less.
7) The ultimate competitive free market economy - Vendor competition in open source is not distorted by the effect of vendor lock in. Open source prevents vendors from actively and purposefully using closed code to maintain their advantage over clients. Vendor competition encourages fair support pricing, great customer service and innovation. It also provides the consumer with a way out if the vendor goes out of business or is not responsive. Open source is a simple survival of the fittest business ecosystem which is driven and focused by evidence based improvement of both health quality and costs.
Taken in aggregate, these advantages create strong financial and quality incentives to join cooperative networks and collaborate. This in turn accelerates improvement of safety and quality through best practice sharing and reducing isolated islands of healthcare data.
Our search for an appropriate EHR led us to VistA in 2000, while researching open source alternatives. Unfortunately, at the time it was nowhere near ready for easy deployment outside the Veterans Administration (VA) so we continued to search for a solid EHR in the usual ways, but found the process disappointing. The process is not unlike being detailed by a pharmaceutical representative, so I started wondering what I could learn by comparing the two. Most physicians don't prescribe medicine based upon what the drug representatives tell them. Instead they use an evidence-based approach. This is now an expectation and considered a standard of care in medicine, because evidence-based medicine saves lives. According to the Institute for Healthcare Improvement, nearly one third of all medical errors could be prevented by applying appropriate technology[2]. So applying technology can save as many lives as prescribing aspirin after a heart attack! I began to wonder, is there an aspirin of electronic health records? What does the evidence based literature say about EHR and impact on quality? Is there one in particular that stands out? Shouldn't applying the medical evidence to the choice of HIT be the standard of care since it, like aspirin, can potentially save so many lives? What I found in the literature shocked me.
It turns out that a search of the peer reviewed medical literature shows that the VA VistA EHR system is one of the only EHR systems that has been associated with improved outcomes. By contrast, the literature says almost nothing about proprietary systems and outcomes. Moreover, VA’s costs only went up 0.8% between 1995 and 2004, while Medicare costs increased by over 40%[3].
Once we understood the role of VistA in the VA's transformation and performance our search was over. In addition we also became aware of the CMS VistA Office EHR initiative, the WorldVistA not-for-profit and the efforts to adapt VistA for use outside the VA. This work would ultimately lead to WorldVistA providing a CCHIT version (WorldVistA EHR) licensed under an open source software license. The only open source EHR to achieve CCHIT certification is WorldVistA.... Suddenly the advantages of the open source model would be available using a CCHIT certified VistA clone!
Clinica Adelante’s WorldVistA EHR Implementation Strategy
So after applying evidence-based studies and recognizing the importance of an open source model in healthcare, we chose WorldVistA to do a demonstration project. We developed a relationship with WorldVistA and became a development site during the CMS project. A key contribution our site made was to pilot a full open source platform which included the open source operating system Linux, and the open source database GT.M to further cut licensing costs.
We leveraged and made use of the extensive resources and documentation which the VA makes available through a number of public web sites such as the VistA University training materials. Other examples of leveraging the open source model include:
l modifying an installation checklist found on the VA documentation website for our use to direct our installation efforts
l developing an open source interface to our practice management system (PMS) for registration and scheduling
l integrating test ordering and results reporting with our external reference lab; our providers order labs in WorldVistA EHR and the results return as discrete data directly into WorldVistA EHR
l development of chronic disease registries that allow data to be entered at point of care and reported in many forms including a HIPAA-stripped form for uploading to state and national chronic disease databases
l implementation of real time drug order checks, automated clinical reminders and automated provider alerts
l development of pediatric templates, including state approved EPSDT forms
We formed 4 teams, using our staff and external consultants to help with the work and build buy in, including our key stakeholders early in the process. We hired a clinician to a training role and hired trainers to train him. The preparation phase took 8 months and we went live August 10, 2007 in Surprise, AZ at our busiest clinic.
Outcomes and Costs
Initially, as with any intervention of this magnitude productivity declined... in our case to 50% of our usual level in the first week, but it recovered to 85-90% in six weeks. We are now at 100% productivity at our first site. Our referrals department can now do 10- 15 referrals per hour, compared to only 6 per hour before implementation. We don't lose medical records any more and they are always available for the patient visit when we need them. We lost no staff or providers as a result of the project. Staff immediately loved the system, but the providers only tolerated it at first. Now, no provider desires to return to the old way or to paper charts.
Our registry functions also appear to be very successful. We now have two registries—one for diabetes and another for asthma--configured. Now 100% of qualified patients are selected automatically for entry by the computer. This will allow planned care to be scaleable to 100% of our patients without hiring extra data entry specialists. We will be able to provide outreach and improved chronic disease management to a much larger population of patients. For instance, when we used the registry that required double entry, we were only able to use Wagoner’s model on about 800 diabetics. Now we can use it on all of our patients with Diabetes. That is over 3000 diabetic patients. We will also be able to extend the Chronic Disease Model to other types of chronic disease, like depression, coronary artery disease and hypertension. Eventually, we hope to give every patient their own personal health plan, using the VistA registry technology.
We were very cost conscious with the first implementation. We had no special grants. Our development costs were approximately $19,000 dollars, plus hardware costs. This does not include the salary of the trainer. Nor does it include lost revenue from staff meetings and lowered productivity, or my time as project leader. To achieve this, I spent most of my administrative time, evening and weekends working on the project. It is doubtful that others can expect to achieve what we did with the same budget, nor should it be so difficult to do the “right thing” by patients.
Since the demonstration project, we have also implemented our EHR at another site and also with the (mobile) rural health team. We are developing a 16 week implementation cycle that can be staggered to allow two implementations in different phases. We have started a network with two other community health centers and a small safety net non-federally qualified clinic. Although the demonstration project allowed us to show clinical success and estimate reduced costs compared to proprietary systems, the project has stalled without more funding. Our analysis of sustainable costs show a savings of 30 to 50% over proprietary systems, perhaps more as the network grows larger. Even so, this cost remains out of reach for most offices. Ultimately, we view the EHR as a tool to reduce medical errors, improve patient care and stabilize the costs of healthcare. Developing these strategies is possible with systems like WorldVistA EHR, but are unlikely to co-evolve on their own. Proper planning, adequate funding and well designed incentives are all necessary to drive projects like these forward. In fact, without more funding, we will not be able implement WorldVistA EHR across all our network sites. This network represents a quarter of a million patient visits a year—that is a lot of patients who we could be reaching and whose care we could be improving with health IT but which we cannot, because of lack of funding.
Based on our practical experience, our view is that VistA is hands down the best system available, is the only solution backed by solid scientific evidence to prove it, and costs 50-70% of the costs of comparable proprietary systems. The fact that it is open source and was developed by with taxpayers' money makes it a logical and very affordable choice for a large segment of the US health system.
Health Improvement through health IT and the need for incentives
Health improvement through health information technology is a tough sell to providers in general because it temporarily affects productivity as providers learn how the use the system. Moreover, any cost savings (like less ER visits because of better control of asthma) are realized downstream from the user and tend to accrue largely to the patient and the health care purchaser. Incentives are a very powerful tool to effect change that successful businesses use all the time. In this context, it is the fastest way to increase the rate of provider adoption for health IT.
Incentives certainly could increase the rate of adoption, but just giving incentives for EHR acquisition will not improve quality. Incentives must be tied to quality improvement or reporting clinical measures to have the desired effect. Connecting offices through networks tasked with quality improvement would work. The most innovative approach would be to move completely away from volume based reimbursement to value based pay. Pay for performance is a step in the right direction, but still relies on volume.
However, it is important to note that quality incentives need an adequate HIT infrastructure with enough connectivity and sufficient granularity to report clinical measures at the provider level. This is why as a first step, I believe it is important that provider incentives be tied to the adoption of EHR systems. I believe further that EHR systems should support these important clinical and quality reporting functions.
In addition, a provider might need time and support to get used to the system and learn to use it effectively. This is why I believe provider incentives should encourage network membership. Networks are better prepared than small offices—much less solo practitioners working on their own—to evaluate EHRs for the necessary functions, have the capital to customize them as needed and the expertise to deploy them, secure them and support them. Networks can also better connect with existing HIE, Medicaid transformation grant projects, labs and other ancillary services, etc. Provider support and clinical improvement will be greater with network formation and will also achieve the goals of better connectivity and improved quality.
Myths about VistA and open source applications
Before I conclude, I want to dispel the many myths floating out there about the VistA system and open source applications in general.
Myth #1: the M coding language is too old to be used in a modern healthcare system. This is false and most large proprietary healthcare vendors, Epic for example, use it. There are many innovations taking place outside the VA right now that show the robust and flexible nature of the M based code.
Myth #2: Open source is unfair in a competitive market. Open source stimulates competition unlike proprietary systems whose goal is to lock in users and monopolize the market. Proprietary systems are only in a competitive market until the client signs on the contract line. Then the relationship becomes very lopsided. I have been to many Health Information Conferences and have listened to the best speakers. They always say deciding on a healthcare vendor is like getting married, because it will be a long-term relationship. It is very difficult to change vendors because of vendor lock. Then they talk in the remaining hour about all the “pre-nuptials” you must get because you can't trust any of the vendors. Open source has competition at multiple levels, but primarily on support services and training which are the most important factors in successful and sustainable adoption of a solution. In the case of WorldVistA EHR both large and small companies can compete against each other with the same a high quality, CCHIT system. Large companies are definitely interested, too. For instance, a major US systems integrator has just won the contract to provide all of Jordan's public health system (46 hospitals, 500 clinics) with the WorldVistA EHR. With open source vendor competition, you reduce price, eliminate vendor lock and improve customer service. Open source is a true free market.
Myth #3: The VA code is too expensive to maintain. VistA, under the open source model has flourished. Clinica Adelante was able to fund an extraordinary amount of customization for a moderate amount of money. Moreover, these enhancements are available for other offices for the price of configuration and support. Some of the code done by WorldVistA has found its way back into the VA system. There is an extraordinary opportunity for governmental agencies like the VA and Indian Health Service to work with private businesses and not for profits to further their missions.
Myth #4: Open source applications are more vulnerable to security breaches. Because open source code is transparent, there is a myth that it is insecure. This has not proved true at all. Breaches are often a result of poor coding practices. The transparency of the code demands that peers code to the highest levels. Moreover, it is scrutinized by expert before it is released. The result is clear: Nobody runs anti-viral software on (open source) Linux, nor do they need to. Everybody runs anti-viral on Windows (closed code) and they would be crazy not to. Moreover, with so many eyes looking at the code, more security flaws are found before breach and more quickly corrected, often within hours.
VistA is the aspirin of EHRs
VistA is the aspirin of EHRs and if it was a drug, every provider would prescribe it. But just like generic aspirin, there are no “drug representatives” or lobbyists to sell it. Its effectiveness is clearly supported in the literature, but administrators don't have time to read the literature. So they listen to the sales pitch and the lobbyists. In the healthcare industry, that could cost lives. In healthcare, when lives are at stake, I believe we should hold ourselves to the same standard we hold our physicians and use the evidence whenever possible to evaluate and select technology solutions...not advertising or marketing hype. And that is why Clinica Adelante chose VistA EHR.
[1] Rothman AA, Wagner EH. Chronic illness management: what is the role of primary care. Ann Intern Med 2003;138: 256-61
[2] Crossing the Quality Chasm: A New Health System for the 21st CenturyCommittee on Quality of Health Care in America, Institute of Medicine, Washington, DC, USA: National Academies Press; 2001
[3] Robert A. Petzel, Director, Veterans Integrated Services Network 23, Compelled to Act: it’s called survival, Powerpoint presentation, slide 14, available at http://www.amq.ca/congres2006/pdf/Compelled_to_Act-Robert_Petzel.pdf