United Cerebral Palsy of Greater Birmingham (UCPGB) is a not-for-profit organization funded through a portfolio of private and public sources that include local foundations, companies, individual gifts, state and federal reimbursement for services provided, grants, and the United Way of Central Alabama. Through various programs, UCPGB provides clinical and health-related services to more than 2,000 infants, children, and adults in Birmingham, Alabama, and in ten surrounding counties.
In October, 2006, UCPGB moved to a new adult facility, LINCPoint, which includes a Health and Wellness Center in addition to the services available at the previous site. The LINCPoint facility is somewhat different from the average physician's office. Because of the population's mobility restrictions, most medical care is provided on site. In addition to staff therapists who provide physical, occupational, and speech therapy, UCPGB retains a variety of outside clinicians (family medicine, optometry, dentistry, etc.) to provide specific clinical services at the facility. Although UCPGB currently has a relatively small adult medical practice in terms of number of patients (approximately 100), the director of medical services anticipates a significant increase in their population as special education students from the area school systems graduate and seek treatment from the UCPGB facility.
The LINCPoint facility has a footprint totaling 42,000 square feet, with nearly 8,000 square feet of space allocated for service delivery. Also included in the total working space is an electrical closet, environmental closet, bathroom, physician's office, nurse's office, and copy room.
While a paper-based medical record system had been considered adequate in the previous setting, the director does not believe it will continue to meet their changing information needs in the new location. The existing client population, with its substantial and varied medical problems, coupled with the expected expansion of the client base, makes a transition to an electronic medical record (EMR) system a business imperative. The director's decision to pursue an EMR was also influenced by the environmental disasters in New Orleans and the Mississippi Gulf Coast. The aftermath of Hurricane Katrina taught many healthcare organizations a painful lesson about the fragility of paper medical records.
Cost of Doing Business with Technology
Money remains the single largest barrier to information technology adoption in smaller physician practices. A recent survey found that 62 percent of solo practitioners and 59 percent of group practices with fewer than 10 physicians cited cost as a "major barrier" to implementing technology solutions (Audet et al. 2005). When hardware, software licensing fees, installation, support, and training are all taken into account, a practice can expect to pay anywhere from $15,000 to $50,000 per physician (Audet et al. 2005; Miller, Sim, and Newman 2003) for an EMR application.
Many commercial EMR vendors set their prices based on the number of physicians in the practice or on the number of users. The per-physician or user pricing method is especially problematic for UCPGB. In addition to the one full-time physician, 10 volunteer faculty clinicians (physicians, dentists, optometrists, and doctorate occupational and physical therapists) treat patients at LINCPoint. Each of these volunteer faculty members brings a cadre of residents to each visit. Thus, under this pricing model, the total implementation cost for UCPGB could easily balloon well into the six-figure range.
Money is a key barrier for UCPGB as well. As a not-for-profit organization, the majority of funding comes from community donations with some support from United Way, the parent organization. The organization reports that 93 percent of all funds received are directed to support and improvement of programs and services. Funds raised to build LINCPoint did not include allocation for an EMR. Although technology-based information resources certainly are expected to benefit clients through improved quality and efficiency, UCPGB has no slack financial resources to purchase a fully functional EMR application.
An effective but economical technology solution was required. In addition to limited financial resources, UCPGB also lacks the personnel resources and technical expertise needed to evaluate available products and plan for system conversion. In short, they needed expert assistance to achieve their strategic objectives for information technology to support their clinical services. Physician Innovations, LLC, a Birmingham, Alabama-based consulting company that specializes in the installation and implementation of EMR systems, was selected to meet that need. The president of Physician Innovations, Marc Krawitz, and principal consultant Meg Bruck led the project. Ed Donaldson, a candidate for a master's degree in health informatics from the University of Alabama at Birmingham, assisted with the project as part of degree requirements.
The process used by the consultant team involved several key steps.
Step One: Understanding User Needs
An effective EMR implementation involves much more than purchasing software and installing it on a computer. Off-the-shelf solutions rarely meet the specific needs of an organization, and even when the "best" vendor applications are identified, customization of functionality, reporting, and service are almost always necessary. These customizations are crucial to overall system usability, user productivity, and, most importantly, user acceptance. The implementation team must understand the needs of the users to negotiate and ensure effective customizations. EMR implementations in the outpatient physician setting have a high failure rate, which takes the form of non-use, inadequate use, or outright rejection of the system. While there are many causes, the lack of adequate customization by the implementation team is prominent on the list of suspects.
Customization begins very early in the planning process, and it requires collecting, organizing, and assessing a significant amount of information related to the current process. The consultant team first used a questionnaire to gather this information. Individual meetings with users were held, using the questionnaire to engage users in a dialog rather than having them respond to the questionnaire independently. The interaction helped the consultants understand the struggles, the time wasters, and the "pain points" of the organization's current process for capturing documents and clinical information.
Once the interviews had been completed, the consultants observed the operations of the clinic in person by shadowing the staff while they worked with patients. This step is important because routine tasks that are second nature to the staff may be omitted in the questionnaire process. Other steps may not be done exactly as described. Sometimes, a new situation arises which needs to be handled differently. These direct observations allow the consultants to document how the work is actually done.
This one-on-one interactive approach was intended to avoid the "user and developer syndrome," the commonly-encountered communication barrier between end-users and technical staff (Leffingwell and Widrig 2000). This syndrome, titled by software engineers, derives from the different training, different terminologies, and different work cultures experienced by these two groups. These differences make it difficult for users and technical experts to exchange information in the needs assessment process.
Step Two: Improving the Existing Processes
Many technology consultants use a simple example to explain why process improvement should occur before new technology is introduced-automating a train wreck will not prevent it from happening, but it makes the wreck happen faster. In EMR implementations, typically the process must be improved from two perspectives- the paper flow and personnel work flow.
Process Improvement: The Paper Chase
The infamous paper medical record is a major contributor to inefficient processes in the physician office setting. Each time information is added to a record, more than a dozen steps may be taken. Typically, these steps include removing the record from the file area, dating and inserting treatment sheets, moving the record to the physician's area, writing progress notes, moving the record back to the business area, and finally re-filing the record. The only step that adds information to the actual record is the writing of the progress note. The remaining steps consume about 68 percent of the total process time in a typical physician office setting (Hultman 1996). Many of these steps must be taken if information is accessed from a record. The movement of many records to multiple locations in an office increases the possibility that a record can be lost.
UCPGB's paper chase problem is compounded by the use of many external clinicians. Each of these clinicians keeps their own copy of the record for their research and teaching purposes as well as for subsequent treatment. Consequently, a great deal of copying and updating is required to keep multiple copies of the record current. Record retrieval and re-filing tasks add to the burden of managing paper records. Clearly, this tedious process is not optimal to support the increased clients expected from UCPGB's long-term objective of involving itself in the care of special needs students throughout the county.
An EMR eliminates these labor-intensive steps by storing patient records on a shared computer system instead of on paper. A good EMR will attempt to make the electronic record look like the paper record, which means including the same information in the electronic record as in the paper record. Identifying the sources of the information in the paper record is key. The best source is the paper record itself. The record contains the forms that doctors use to make notes and give treatment orders, and it contains the forms the nurses use for collecting vital signs and administering treatment. There are also paper documents from external sources, such as lab test results, imaging pictures and the corresponding reports, and even records from other doctors. It may not be desirable to add every single piece of paper to the EMR, but all documents need to be identified to make informed decisions about what to include and what to leave out.
Just as important as the information that goes into the record is the information that comes out of the record. What kinds of reports are created in the office? Is there any regulatory reporting that needs to be generated from the new EMR? Sometimes, physicians need to do separate reporting on certain diagnoses, such as diabetes or breast cancer. Doctors offices are frequently asked to provide records to new doctors. How will this be accomplished?
If the practice has an information-sharing relationship with another organization, it may be appropriate to create an electronic interface with that organization's EMR. If no such relationship exists, or if there are no other EMRs to interface with, then standardized reports should be created to be printed and faxed or mailed on request.
Process Improvement: The People Chase
With the focus and attention on creating a paperless office, other steps in the workflow are often overlooked. EMR implementation is not just about reproducing the paper chart on the computer. The software is centered on the workflow, which includes how doctors and other staff communicate and share information. The EMR can improve the efficiency of communication if it is designed for that purpose. Therefore, it is important to document the current communication process during the data gathering phase.
The consultants learned that UCPGB has a very informal work atmosphere. It is a friendly place, and everyone enjoys working there. But frequent interruptions, even if they are friendly, are disruptive to one's concentration on a job task. The staff and patients move around a lot, and business may well be conducted in the hallways. Staff frequently enter each other's offices to ask questions or give updates during their encounters with individual patients. This approach has become a routine way of doing business.
For example, if the nurse examines a patient and notes that she needs to see a dentist, the nurse typically will go into the medical director's office and tell him verbally so he can sign off on the order. The medical director stops his work to write down the information to process an order at a later time. The message is important, but not urgent. The nurse is conveying information when it is most convenient for her, but not at the most convenient time for the physician. A similar situation might occur many times during the day.
An EMR system can be configured to automate this type of communication. The nurse examines the patient's teeth and makes an entry in the electronic record, noting the need for a dental referral. She requests a counter-signature from the medical director, which causes the note to appear on his notification list the next time he logs into the system, along with other items that need his attention. With one click, he reads the nurse's note and electronically signs the order.
This is a simple example of process improvement. It is not always possible to reach maximum efficiency or to create an entirely automated, paperless environment. Such achievements often are too overwhelming, too drastic, or too expensive. In most cases, significant improvement can be made by finding the biggest time-wasters and focusing energy on removing those barriers. Perhaps you can identify a few steps in a process that are done, but nobody remembers why. Small-scale changes can leave the staff with work routines that are familiar and comfortable, but the changes make the process work better or quicker. In most cases, the staff attribute the process improvements to the technology rather than to improved work processes.
Step Three: System Selection and Configuration
The criteria used for selection of an EMR system for UCPGB were simple-adequate functionality to meet user needs at an affordable price. After determining the documentation needs and improving the work processes, the consultants recommended implementing VistA, the EMR developed by the Veteran's Health Administration approximately 20 years ago. VistA was developed as free and open source software (FOSS) system, which means the application and the source code are public and available to anyone. Any individual or organization with the necessary technology skills can make custom changes to the application. A FOSS convention is that users share improvement, modifications, and tips with the global open source community to improve the utility of the software for all.
The VistA EMR covers all aspects of the health system, from scheduling and patient registration to electronic order entry and electronic clinical documentation. Several VistA options are available to users. The option chosen for UCPGB was VistA Office, which runs the client component on Microsoft Windows. The server component can run on Windows or Linux; UCPGB chose to use open source Linux, which is considered less prone to virus attacks and security breaches. Linux is rated as low cost, high performance, and high reliability.
Step Four: Prototyping
Once the information is gathered and work processes have been improved, it is time to start the process of configuring and customizing the chosen system. This initial version is known as a prototype. It is not intended as a final product, but rather something to show users to elicit feedback about how it will work in their organization. Rapid prototyping is an iterative process in which the system is modified, feedback is sought, and further modifications are made. This cycle occurs many times until, eventually, a nearly ideal system results. Prototyping saves money and results in a more usable system.
An important part of developing the prototypes was ensuring service integration. UCPGB provides comprehensive healthcare services to its patients that cross multiple specialties, including family medicine, physical and rehabilitative medicine, physical and occupational therapy, optometry, dentistry, and psychiatry. The medical director seeks an integrated healthcare approach in which these specialists cooperate, collaborate, and communicate to provide non-fragmented care.
The necessity for an EMR is magnified in scenarios like this-each clinician must see a complete picture of a patient's health status. For example, medication changes may occasionally have a deleterious effect on a patient's functional capacity in other areas (i.e., speech or movement). With paper records, medication changes may not be communicated to other providers efficiently. With an EMR, all providers have instant access to all current medications. Electronic communication facilitates discussion and questioning among providers to track down potential causes of such problems.
Progress note templates, order sets, and quick orders must be defined for each specialty and in some instances for individual clinicians. Additionally, scheduling parameters for each specialty are different. For example, occupational therapy is typically not scheduled at all, but is given on an as-needed basis. Speech therapy is given in a group therapy context, while many of the other specialists are onsite on a weekly, monthly, or quarterly basis.
Optometry and dentistry presented some particular difficulties. Although templates can be defined to document care for these specialties, VistA does not provide specialized support such as clickable teeth or the ability to draw on an eyeball to indicate retinal deterioration. However, most general purpose EMRs would have the same limitation-only specialized dental and optometry vendors would provide such support. To handle this situation, some document scanning was needed. In certain situations, clinicians draw diagrams on paper and scan in the drawing. VistA can store and display this information as part of the medical record. In general, EMRs should store structured information as much as possible. Realistically, there will be situations in which information needs to be scanned in. Other examples might be certain lab tests, wave forms, etc.
Step Five: Customizing
A well-configured EMR system should be as "point-and-click" as possible. Typing text into orders is occasionally necessary, but when possible, users should instead use their mouse or tablet pen to choose from a carefully honed set of options. For example, each clinical specialty at UCPGB has categories of frequently written orders. A physician specialist will periodically visit the facility and assess the patients. Orders are written by the physician for therapists and other staff to carry out. A customized set of standard orders was developed for each specialty area at UCPGB.
Medication and diagnostic orders may also be customized for efficiency. VistA has two ways to make ordering easier. The first involves order sets, which are a collection of related orders frequently given together. For example, if a patient complains of a severe sore throat, a physician might order both a rapid strep test and a mono test. These could be combined together into an order set. The second technique involves quick orders, which are predefined individual orders that are frequently prescribed. As an example, a high percentage of UCPGB patients are given a prescription for transdermal scopolamine patches which are used to control excessive drooling. Consequently, it makes sense for UCPGB to have a quick order for this drug.
Progress note templates are the electronic equivalent of paper forms with check boxes for different patient diagnoses and conditions. The utilization and customization of templates was essential for adequate clinician productivity at UCPGB. Without templates, a clinician would need to create the text of progress notes each time, which can be very time consuming. Many different template come with VistA Office and users may create their own as well.
A particular problem with paper records at UCPGB involved documenting group therapy. Two full-time speech therapists perform much of their work in group sessions with thee to eight patients at a time. Entering progress notes to document their care is a monumental undertaking since the manual process discussed previously is multiplied by a factor of up to eight! VistA handles this problem elegantly by employing its group notes feature. In this scenario, a clinician can create a progress note and select multiple patients. The progress note is automatically copied to the charts of all selected patients. The clinician has the option to customize each patient's copy if special considerations warrant such action. An order-of-magnitude productivity improvement over traditional paper-based processes is easily possible in this situation.
Step Six: Populate and Implement
Once the customized system works in the test environment, it is time to move to real time. The customized templates are approved, the order sets are as complete as possible, the drug lists have been updated, and the CPT (current procedural terminology) codes have been uploaded. The software is installed on every workstation, the tablets are configured, and even all the users and logins are created. It is time to move into full implementation and go "live!" However, to be immediately useful for continuity of care, UCPGB's existing patient date must be incorporated into the system.
Virtually all EMR software comes with some EDI (electronic data interchange) capability. Some may use custom formatting or HL7 messages, while other use comma-separated values or fixed-width batch files. VistA Office has an HL7 interface as well as a batch patient creation feature so that multiple patients can be uploaded in one file.
A file upload assumes the information is already available electronically. If the implementation is a conversion, such as from billing or practice management software, then the necessary data exists electronically. It may not be easy to create the interfaces between two software systems, but it can be done and is worth the hours and dollars spent to get it right. The alternative is to re-key all of the relevant data from paper records.
Unfortunately, UCPGB had no electronic records; they were totally paper-based. Medical records were all handwritten, appointments were kept in notebooks, and bills were sent to Medicaid on paper forms. The only electronic document was a Word file listing patients, addresses, Medicaid numbers, and other contact information that served as the patient index. This document could have been converted to a usable batch file, but not all patients had the same information listed, the format was not consistent, and some information was missing. Therefore, the best option was to manually enter the data.
It took three people two weeks to properly load ninety patients into VistA Office. In retrospect, it may have been easier to reformat the Word document to a .csv file and work through the VistA Office batch-loading feature.
Once the patients' identification data were entered into the system, it was time to address their historical and continuing clinical data. The therapists were onsite daily and had ready access to their existing extensive paper charts. A "history" of therapy information could be generated relatively quickly with daily entries and ongoing therapy sessions. Their ability to refer to the paper records while building the electronic record in real time eliminated the need to enter all their historical data manually.
Unfortunately, this was not an option for the medical specialist who came on site infrequently and maintained records in multiple locations. Clinicians cannot benefit from an electronic chart without usable history data for each patient. Migrating data from one EMR to another would be a relatively easy task using HL7. However, the UCPGB records were strictly paper based, and manual data entry was required. Data entry of this nature must be performed by someone who is familiar with the terminology and understands the significance of the data. In this case, that "someone" would be a doctor or nurse.
The medical services director volunteered to enter data from as many patient records as possible while the center was closed during a holiday period. After reviewing the available templates and the data in each record, he used the primary care new patient template to enter the complete history and physical. He entered all the available information from the most recent exams with each provider. After this initial data entry, more detailed patient histories were entered prior to scheduled visits for the respective patients.
This approach allowed the facility to spread the workload over a period of time. An alternative approach would have been to use a hybrid model, in which clinicians use both the EMR and the paper record for about three visits. During after each visit, the new progress notes, orders, and pertinent history are entered into the EMR. The paper record is used for reference only, and its use should quickly diminish. Either way, it cannot be overemphasized that the entry of medical history is a time-consuming and labor-intensive process.
Step Seven: User Training
EMRs are complex systems, and most users need extensive training to use the system efficiently and to maximize its functionality. Training for UCPGB started with a group demonstration of the software to give the clinicians a feel for the basic capabilities of the system. This was followed by one-on-one training sessions with each clinician. A Physician Innovations staff member would sit next to a UCPGB clinician while they navigated the system to provide guidance and tips. The two individuals would frequently switch places at the computer terminal to demonstrate a technique, ask questions, or tweak the system when necessary. This one-on-one approach was pivotal to the success of the project since it enabled questions and concerns to be addressed in real-world scenarios and in real time. Additionally, since order sets and templates are specialty-specific, the training was focused specifically to the user's needs. Training sessions involving individuals from all specialties would have been less efficient.
Another aspect of training involved using tablet personal computers (PCs). Wireless tablet PCs have substantial workflow advantages by bringing the EMR directly to the point of care. However, tablets themselves have a learning curve, particularly those which lack a keyboard (slate form). By contrast, "convertible" tablets have a keyboard that can be twisted and rotated to hide it. Slate tablets are lighter and thinner, while convertible tablets may be preferred by users who are not comfortable handwriting on a touch sensitive computer screen. Users should try both options prior to purchase. UCPGB chose the slate model, but also purchased a docking station and a Bluetooth wireless keyboard and mouse. This configuration gives users the best of both worlds-the familiarity of a keyboard and mouse along with the flexibility and mobility of a slate tablet.
Step Eight: Evaluation and Planning for Maintenance
Once the system is fully operational, the final implementation step is to evaluate the system's functionality and usability, and to assess user satisfaction with the system. After any final adjustments or tweaks are made, a plan for ongoing maintenance of the system should be established. This plan should address vendor upgrades, problems encountered by the users, and any eventual linkage with or migration to another system. In light of VistA's FOSS status, monitoring user modifications released in the global community should be addressed as well.
Audet, A. M., M. M. Doty, J. Shamasdin, and S. C. Schoenbaum. 2005. Physicians' Views on Quality of Care: Findings From The Commonwealth Fund National Survey of Physicians and Quality of Care. Report no. 823 (May). New York: Commonwealth Fund.
Hultman, J. A. 1996. St. Anthony's Reengineering the Medical Practice: Profit Through Efficiency in a Medical Office Environment. Eden Prairie, MN: St. Anthony Publishing.
Leffingwell, D. and D. Widrig. 2000. Managing Software Requirements: A Unified Approach. Indianapolis, IN: Addison-Wesley.
Miller, R. H, I. Sim, and J. Newman. 2003. Electronic Medical Records: Lessons from Small Physician Practices. California HealthCare Foundation iHealth Report. [Online article; retrieved 04/17/08.] http://www.chcf.org/documents/ihealth/EMRLessonsSmallPhyscianPractices.pdf
The migration from paper record system to an EMR system in a clinical/hospital setting is truly a daunting one. The case study may make it sound rather simple and almost easy, but technology migrations are never simple or easy, and even after the new system is in place and in use, the work is not over, since systems, especially new systems, must be maintained. That said, the attached case study is more of a system migration checklist than a detailed history of UCPGP's particular EMR story. By this, I mean that while there are details about the choices UCPGP made during the migration process, there is little detail about the basis for those decisions and worse still, almost no color on the effects of those decisions on the implementation process. However, there are only two questions posed for this case study, and they are general enough. One question goes to how to measure user satisfaction and ongoing training for the new system, and the second question goes to what changes could have been made in both the design and implementation phases of UCPGP's migration, as well as to interoperability.
USER SATISFACTION AND ONGOING TRAINING
In this particular case, it is incumbent on hospital management and the individual clinicians to ensure that the new EMR system is maintained to its highest level of efficiency - both for UCPGP's patients and for the smooth functioning of the facility. Because the majority of people are uncomfortable with massive changes (in any area of work and life), sweeping changes (such as the migration from paper-based record keeping to an EMR system) are always challenging. User satisfaction and overall system health can often be accomplished by allocating a significant percentage of migration resources to "training". If I know how to use the system, I am more likely to use it correctly. If I can use it correctly, more often than not, I am more likely to feel confident about what I am doing (both my job and the system I am using in order to do my job). If I am more confident at work, I am, in general, more satisfied. Thus, "training" is a great way to introduce ...
This solution discusses aspects of a hospital EMR implementation, with a focus on user satisfaction, ways the implementation could have been improved and how interoperability figures into the project process.