Did Your EHR Make the Grade?
A successful EHR implementation doesn't happen overnight. The planning and implementation is complex and fraught with opportunities for practice mismanagement. Not surprisingly, not all implementations are successful. Equipped with a "failure is not an option" mentality, how do providers move forward and recover in the face of IT failures?
It's difficult to pin down an accurate number for an average EHR implementation failure rate. A litany of data cites failure rates hovering around 50 percent, while another report found that 19 percent of EHRs are left uninstalled. One 2009 report by health IT consulting firm AC Group puts the rate as high as 73 percent due to usability frustrations.
Despite the varying figures, EHR failures are a tough reality. Organizations are naturally reluctant to discuss these failures. No one wants to look organizationally weak, especially through public exposure.
But, some thought leaders are encouraging providers to share their failures for the sake of greater healthcare progress. "Every industry other than health IT recognizes that real learning comes from failures," says Jonathan A. Leviss, MD, chief medical officer at the Rhode Island Quality Institute (RIQI) in Providence, R.I., who maintains that these lessons create lasting impressions.
To him, a lecture outlining four critical factors leading to a health IT failure would be more memorable and useful than a lecture advocating four success factors. "When I go back to my organization to take on a health IT project, I want to make sure that I don't do the same four things that led to a failure, even if I only follow three of the four success factors."
Constantly evaluating the EHR initiative is critical, according to Leviss. "Part of project management is regularly asking 'What's going wrong now?'" he says. "It could range from exceeding the budget to vendors delivering software missing critical functionality to not having the right clinicians involved in the project."
Limited functionality, along with a host of other issues, led Michael J. West, MD, PhD, clinic director at the Washington Endocrine Center in Washington, D.C., to fire his first EHR vendor. Opening his practice in 2009, West bought an EHR system for $10,000 and put a server with thin-client terminals into his office after two weeks of researching systems. Almost immediately, technical issues became a problem. For example, the system was riddled with broken software links. When one piece of the software was fixed, other parts would cease functioning and require repair.
"Training was never completed because the system never worked properly," recalls West. With the intention of training via remote desktop sessions, improperly installed software wouldn't allow for adequate, complete EHR training.
The system seemed to be partially broken every day, with the vendor continuously promising to fix the system. The patient notes were formatted awkwardly, says West, and weren't intuitive to how a physician actually practices medicine. "For a patient's family history, if a mother died at 65 from a heart attack, I had to know the ICD-9 code for heart attack and the drop-down lists seemed unending," says West. "Does the patient drink alcohol? How many ounces a day? What type of alcohol? Every week of the month? Every day? The system knew the billing language but the documentation wasn't meaningful. I didn't use certain parts of the system and mostly used the free text box for documentation."
Part of the problem, West adds, is that he believes the vendor didn't recommend the correct hardware for his practice to purchase for maximum functionality. The vendor provided two specification sheets, one was emailed to an IT consultant and another to West. When West and the IT consultant compared the specification sheets, the recommended hardware differed. Despite the vendor dismissing the question about the server he had purchased, West never received a good explanation for why they recommended purchasing the two listed servers.
After three months of being told the system could be made more customizable with little results, West fired the vendor in February 2010. He asked for a full refund claiming product misrepresentation, but ultimately settled for a partial refund.
Because of his experience, West has vowed never again to buy in-house EHR technologies, opting instead for web-hosted software to be able to log in and go to work. "We plan on renting from now on and if a service doesn't work, we can cut the cord right then and there. I wouldn't invest in any more hardware." He is currently using no-cost, web-hosted software for electronic charting that employs free text, so he can write a patient note in whichever style and language he desires. He also recently unplugged his server.
Change is hard and health IT executives know that changing how physicians conduct business can be trying. One of the first examples of a significant IT failure dates to 2003. A Los Angeles Times article reported that approximately 400 physicians at Cedars-Sinai Medical Center in Los Angeles confronted hospital administrators leading to the cancelation of a computerized physician order entry (CPOE) system, citing the system's safety issues (Charles Ornstein. LA Times Jan. 22, 2003). The CPOE system, according to the article, is rumored to have cost the healthcare system $34 million. Cedars-Sinai officials did not verify that figure to Clinical Innovation + Technology, but offered this statement:
"Implementing CPOE is both difficult and complicated. Cedars-Sinai's first attempt at CPOE failed for many reasons. Among those was our need to learn how to build and deploy CPOE since we started so early. We recently have redeployed CPOE at Cedars-Sinai to strong reviews."
Leland A. Babitch, MD, MBA, CMIO at Detroit Medical Center, had similar CPOE woes. In 2004, prior to his tenure, Detroit Medical attempted to implement CPOE in psychiatric and rehabilitation units, which were chosen for their limited scope of practice. With the IT department handling the tool's design, implementation and training, Babitch says that the initial go-live went poorly. "The order catalog was not intuitive nor well-designed," he recalls. "There was inadequate training and support for clinicians, and the launch wasn't integrated with nursing documentation or other key functionality such as an electronic medical administration record."
From the beginning, nurses complained that the new tool was more a beast of burden than a bastion of best practices. While there was no evidence of patient harm for a wrong medication or a missed dose during this time, Babitch notes, user dissatisfaction quickly led Detroit Medical to pull the CPOE plug.
While exhausting in the short run, learning from implementation failures can guide long-term success. Mason General Hospital & Family of Clinics in Shelton, Wash., used its shortcomings on the hospital side to guide EHR implementation in the emergency department while simultaneously retooling the existing hospital system. Dean Gushee, MD, Mason General's medical director, says in 2006, the organization was ready to roll out an EHR system throughout the hospital but the usability didn't cut the mustard. Nurse documentation wasn't intuitive and unanticipated costs added up.
"As we began to roll out the system, we found it took an enormous amount of staff time because they were provided with a framework to build screens that worked within the flow of individual departments. That's a huge amount of nursing and physician time to design and build these screens so documentation occurs," says Gushee. "Those costs don't go away because it is five years later, and we're still redesigning screens as our work processes change on the hospital side."
Gushee also found that multiple modules don't often talk well together. For example, the plan was to roll out CPOE in the operating room before going facility-wide. However, Gushee found an unrecognized problem on the vendor side: if you move a patient from post-anesthesia to recovery out to the hospital floor, all of the orders get canceled regardless of who wrote them, even though some orders might need to be pushed forward to the next caregiver. Because this is a large patient safety concern, the organization went back to paper order sets. "That's truly frustrating when vendors have developed this software but they don't understand the intricacies of the clinician workflow. It feels like it was designed by engineers, not clinicians," Gushee says.
Looking to build a better emergency department system, Gushee in June 2006 implemented an EHR system built and designed by clinical people and found the number of patients left without being seen decreased 10 percent almost immediately. In addition, emergency department revenues improved 20 percent and door-to-doctor time declined 60 to 40 minutes overall, Gushee notes. Currently, Mason General is evaluating new system options.
In the second round of CPOE implementation at Detroit Medical, Babitch similarly found that a clinician-led design and implementation drove successful adoption. In 2005, Detroit Medical reinitiated the CPOE process. With a new CEO at the helm, a CPOE initiative was revamped as a means to stay competitive in the competitive Detroit provider market. "The CEO believed that our hospitals could differentiate themselves by providing high quality and safety-driven care through an advanced EHR system," Babitch says.
Using a clinician-based multidisciplinary approach, Detroit Medical rolled CPOE out to eight hospitals from April 2006 to May 2007, a record scope and pace for that time.
After the successful implementations of CPOE, nursing documentation, closed-loop medical scanning and a new pharmacy system, Detroit Medical has never looked back nor slowed down. "We have continued to upgrade and implement new technologies throughout our hospitals. We have new smart rooms at two of our sites, extensive use of electronic physician documentation, and have attained HIMSS Stage 6 at all of our hospitals with a goal of being paperless in the next few years," says Babitch.
From a marketing perspective, EHR vendors are going to tout the bells and whistles of their systems. But can you trust the verbiage? West attributes the purchase of his system to the sales pitch. "There was nothing special about the system, but it was promised to function how we wanted it to and we were told notes could be customized at our request. We assumed it would be a good system that wouldn't require much customization."
Providers considering the purchase of an EHR should ask for a sample patient note from the vendor, West recommends. "Ask for a PDF of what the note looks like on a printed page." He also warns not to immediately jump into an agreement with a vendor just because their software is certified. The vendor that he fired was certified by the Certification Commission for Health IT, so "it's a buyer's beware market."
Babitch warns that certain vendors have stability problems at the moment. "Would you choose to go with someone if there's a possibility the system will fall apart?" he poses. While Babitch offers no answers, he notes it's an important question that CMIOs should ask themselves. While implementing an EHR for the children's hospital, Detroit Medical debated whether to choose a best-of-breed or single-source system. Compromises associated with the choice of a best-in-breed system have led to delayed go-lives with less functionality than originally intended," he says. "It is impossible to know if progress would have been better with a single-source vendor or a company with better internal stability."
As an advocate of a best-of-breed approach from a clinician standpoint, Gushee says EHR systems shouldn't sacrifice end-user experience for the mantel of interoperability. "I'm a firm believer of interoperability, because that's the direction we have to go, but some institutions are pushing interoperability hard, superseding the end-user experience. Usability has to matter."
As organizations begin to adopt and adapt to new technology, it will be necessary for failures to be discussed, whether in the public or private domain, to create and ensure best practices are moving forward into a new clinical IT world.
It's difficult to pin down an accurate number for an average EHR implementation failure rate. A litany of data cites failure rates hovering around 50 percent, while another report found that 19 percent of EHRs are left uninstalled. One 2009 report by health IT consulting firm AC Group puts the rate as high as 73 percent due to usability frustrations.
Despite the varying figures, EHR failures are a tough reality. Organizations are naturally reluctant to discuss these failures. No one wants to look organizationally weak, especially through public exposure.
But, some thought leaders are encouraging providers to share their failures for the sake of greater healthcare progress. "Every industry other than health IT recognizes that real learning comes from failures," says Jonathan A. Leviss, MD, chief medical officer at the Rhode Island Quality Institute (RIQI) in Providence, R.I., who maintains that these lessons create lasting impressions.
To him, a lecture outlining four critical factors leading to a health IT failure would be more memorable and useful than a lecture advocating four success factors. "When I go back to my organization to take on a health IT project, I want to make sure that I don't do the same four things that led to a failure, even if I only follow three of the four success factors."
Constantly evaluating the EHR initiative is critical, according to Leviss. "Part of project management is regularly asking 'What's going wrong now?'" he says. "It could range from exceeding the budget to vendors delivering software missing critical functionality to not having the right clinicians involved in the project."
Another One Bites the Dust
Limited functionality, along with a host of other issues, led Michael J. West, MD, PhD, clinic director at the Washington Endocrine Center in Washington, D.C., to fire his first EHR vendor. Opening his practice in 2009, West bought an EHR system for $10,000 and put a server with thin-client terminals into his office after two weeks of researching systems. Almost immediately, technical issues became a problem. For example, the system was riddled with broken software links. When one piece of the software was fixed, other parts would cease functioning and require repair. "Training was never completed because the system never worked properly," recalls West. With the intention of training via remote desktop sessions, improperly installed software wouldn't allow for adequate, complete EHR training.
The system seemed to be partially broken every day, with the vendor continuously promising to fix the system. The patient notes were formatted awkwardly, says West, and weren't intuitive to how a physician actually practices medicine. "For a patient's family history, if a mother died at 65 from a heart attack, I had to know the ICD-9 code for heart attack and the drop-down lists seemed unending," says West. "Does the patient drink alcohol? How many ounces a day? What type of alcohol? Every week of the month? Every day? The system knew the billing language but the documentation wasn't meaningful. I didn't use certain parts of the system and mostly used the free text box for documentation."
Part of the problem, West adds, is that he believes the vendor didn't recommend the correct hardware for his practice to purchase for maximum functionality. The vendor provided two specification sheets, one was emailed to an IT consultant and another to West. When West and the IT consultant compared the specification sheets, the recommended hardware differed. Despite the vendor dismissing the question about the server he had purchased, West never received a good explanation for why they recommended purchasing the two listed servers.
After three months of being told the system could be made more customizable with little results, West fired the vendor in February 2010. He asked for a full refund claiming product misrepresentation, but ultimately settled for a partial refund.
Because of his experience, West has vowed never again to buy in-house EHR technologies, opting instead for web-hosted software to be able to log in and go to work. "We plan on renting from now on and if a service doesn't work, we can cut the cord right then and there. I wouldn't invest in any more hardware." He is currently using no-cost, web-hosted software for electronic charting that employs free text, so he can write a patient note in whichever style and language he desires. He also recently unplugged his server.
Changing technology begets change management
Change is hard and health IT executives know that changing how physicians conduct business can be trying. One of the first examples of a significant IT failure dates to 2003. A Los Angeles Times article reported that approximately 400 physicians at Cedars-Sinai Medical Center in Los Angeles confronted hospital administrators leading to the cancelation of a computerized physician order entry (CPOE) system, citing the system's safety issues (Charles Ornstein. LA Times Jan. 22, 2003). The CPOE system, according to the article, is rumored to have cost the healthcare system $34 million. Cedars-Sinai officials did not verify that figure to Clinical Innovation + Technology, but offered this statement: "Implementing CPOE is both difficult and complicated. Cedars-Sinai's first attempt at CPOE failed for many reasons. Among those was our need to learn how to build and deploy CPOE since we started so early. We recently have redeployed CPOE at Cedars-Sinai to strong reviews."
Leland A. Babitch, MD, MBA, CMIO at Detroit Medical Center, had similar CPOE woes. In 2004, prior to his tenure, Detroit Medical attempted to implement CPOE in psychiatric and rehabilitation units, which were chosen for their limited scope of practice. With the IT department handling the tool's design, implementation and training, Babitch says that the initial go-live went poorly. "The order catalog was not intuitive nor well-designed," he recalls. "There was inadequate training and support for clinicians, and the launch wasn't integrated with nursing documentation or other key functionality such as an electronic medical administration record."
From the beginning, nurses complained that the new tool was more a beast of burden than a bastion of best practices. While there was no evidence of patient harm for a wrong medication or a missed dose during this time, Babitch notes, user dissatisfaction quickly led Detroit Medical to pull the CPOE plug.
Testy Times
While exhausting in the short run, learning from implementation failures can guide long-term success. Mason General Hospital & Family of Clinics in Shelton, Wash., used its shortcomings on the hospital side to guide EHR implementation in the emergency department while simultaneously retooling the existing hospital system. Dean Gushee, MD, Mason General's medical director, says in 2006, the organization was ready to roll out an EHR system throughout the hospital but the usability didn't cut the mustard. Nurse documentation wasn't intuitive and unanticipated costs added up. "As we began to roll out the system, we found it took an enormous amount of staff time because they were provided with a framework to build screens that worked within the flow of individual departments. That's a huge amount of nursing and physician time to design and build these screens so documentation occurs," says Gushee. "Those costs don't go away because it is five years later, and we're still redesigning screens as our work processes change on the hospital side."
Gushee also found that multiple modules don't often talk well together. For example, the plan was to roll out CPOE in the operating room before going facility-wide. However, Gushee found an unrecognized problem on the vendor side: if you move a patient from post-anesthesia to recovery out to the hospital floor, all of the orders get canceled regardless of who wrote them, even though some orders might need to be pushed forward to the next caregiver. Because this is a large patient safety concern, the organization went back to paper order sets. "That's truly frustrating when vendors have developed this software but they don't understand the intricacies of the clinician workflow. It feels like it was designed by engineers, not clinicians," Gushee says.
Looking to build a better emergency department system, Gushee in June 2006 implemented an EHR system built and designed by clinical people and found the number of patients left without being seen decreased 10 percent almost immediately. In addition, emergency department revenues improved 20 percent and door-to-doctor time declined 60 to 40 minutes overall, Gushee notes. Currently, Mason General is evaluating new system options.
In the second round of CPOE implementation at Detroit Medical, Babitch similarly found that a clinician-led design and implementation drove successful adoption. In 2005, Detroit Medical reinitiated the CPOE process. With a new CEO at the helm, a CPOE initiative was revamped as a means to stay competitive in the competitive Detroit provider market. "The CEO believed that our hospitals could differentiate themselves by providing high quality and safety-driven care through an advanced EHR system," Babitch says.
Using a clinician-based multidisciplinary approach, Detroit Medical rolled CPOE out to eight hospitals from April 2006 to May 2007, a record scope and pace for that time.
After the successful implementations of CPOE, nursing documentation, closed-loop medical scanning and a new pharmacy system, Detroit Medical has never looked back nor slowed down. "We have continued to upgrade and implement new technologies throughout our hospitals. We have new smart rooms at two of our sites, extensive use of electronic physician documentation, and have attained HIMSS Stage 6 at all of our hospitals with a goal of being paperless in the next few years," says Babitch.
Talk to Your Vendor
From a marketing perspective, EHR vendors are going to tout the bells and whistles of their systems. But can you trust the verbiage? West attributes the purchase of his system to the sales pitch. "There was nothing special about the system, but it was promised to function how we wanted it to and we were told notes could be customized at our request. We assumed it would be a good system that wouldn't require much customization." Providers considering the purchase of an EHR should ask for a sample patient note from the vendor, West recommends. "Ask for a PDF of what the note looks like on a printed page." He also warns not to immediately jump into an agreement with a vendor just because their software is certified. The vendor that he fired was certified by the Certification Commission for Health IT, so "it's a buyer's beware market."
Babitch warns that certain vendors have stability problems at the moment. "Would you choose to go with someone if there's a possibility the system will fall apart?" he poses. While Babitch offers no answers, he notes it's an important question that CMIOs should ask themselves. While implementing an EHR for the children's hospital, Detroit Medical debated whether to choose a best-of-breed or single-source system. Compromises associated with the choice of a best-in-breed system have led to delayed go-lives with less functionality than originally intended," he says. "It is impossible to know if progress would have been better with a single-source vendor or a company with better internal stability."
As an advocate of a best-of-breed approach from a clinician standpoint, Gushee says EHR systems shouldn't sacrifice end-user experience for the mantel of interoperability. "I'm a firm believer of interoperability, because that's the direction we have to go, but some institutions are pushing interoperability hard, superseding the end-user experience. Usability has to matter."
As organizations begin to adopt and adapt to new technology, it will be necessary for failures to be discussed, whether in the public or private domain, to create and ensure best practices are moving forward into a new clinical IT world.
| EHR Failure Usually Involves One Or More of These Four Issues: |
| 1. Technical EHR implementation failures, where the wrong hardware/software combination was installed, or where there were issues with wireless connectivity; 2. Financial failures, where the expected EHR ROI wasn’t realized, or the costs were significantly more than expected; 3. Software incompatibility issues, where the EHR system didn’t interface with an existing medical practice management system; and 4. People-related issues, where some physicians or staff members avoid training or simply refuse to use the EHR system. This can leave a medical practice with the worst of both worlds: paper charts for some physicians and patients, and electronic records for others. Source: Excerpted from “The Top Four Reasons for EHR Implementation Failure” by Jim Hook, MPH, director of consulting for The Fox Group. |