Health IT Best Practices: Innovation In the Trenches
CMIO's second annual Health IT Top Trends Survey, conducted last summer, found that the top health IT implementation priorities of clinical decision support and quality improvement/reporting tools tied for the top spot, followed by EHRs. These priorities probably surprised no one considering the timing of the HITECH Act. Organizations across the U.S. are working to meet the meaningful use criteria of EHRs, but few facilities have mastered these system implementations, let alone the optimization that HITECH was designed to foster.
"Clinical decision support (CDS) is an area in its relative infancy in the electronic world and definitely a vital area to focus on going forward," says Jeffrey S. Rose, MD, vice president of clinical excellence informatics of Ascension Health in St. Louis. "CDS hasn't been [implemented or utilized] thoroughly or very well in most places yet. The hope is that after the first phase of meaningful use, people will turn their attention to the whole idea of making improvements at the point of care."
In the post-implementation world, the CMIO's role will be to continue to show the value of these systems, according to Christopher Longhurst, MD, CMIO of Lucile Packard Children's Hospital (LPCH) at Stanford University in Palo Alto, Calif.
But will the HITECH Act and the requirements it mandates really improve patient care? Jason Hwang, MD, MBA, executive director of healthcare at Innosight Institute, a nonprofit think tank in Mountain View, Calif., and co-author of The Innovator's Prescription: A Disruptive Solution for Health Care, isn't so sure.
The HITECH Act and subsequent meaningful use requirements are well-intentioned, Hwang says, but simply implementing computerized physician order entry (CPOE) and EHRs is not enough. "The real benefits of health IT are going to result from new capabilities like data analytics, enabling team-based care, data portability and interoperability, as well as the incorporation of many new streams of data via wireless devices and sensors. Whereas CPOE is merely a modernization of what we do already, these other benefits are truly novel and will bring far more impactful changes to how we can deliver affordable, quality care."
Many facilities have expanded their vision far beyond implementation and developed ways to leverage their health IT for tangible improvements in efficiency and patient outcomes. After all, as Hwang says, "many patients are already searching online for answers to their medical questions before they ever see their doctor. The real challenge will be in convincing healthcare professionals to adopt health IT with the same fervor."
Time will tell as we move toward the next stages of meaningful use whether these initiatives translate to meaningful improvements. CMIO talked to Rose, Longhurst and other leaders at facilities on the leading edge of health IT innovation.
Landa recommends keeping in touch with the "idea people" during the planning phases of CPOE and electronic records. "By including them early in the discussion, we can help them with those ideas without a tremendous amount of work added to our plates."
Alameda County will undergo EHR implementation over the next year. To prepare, the organization began using a relatively simple text-based order entry system that allowed for electronic creation of order sets. Before, orders were simply printed on paper. "We still create paper order sets but we added the decision support you can get from an electronic process and removed order transmittal."
These "smart orders" require answers to certain questions before users can print. The questions are relatively intuitive and easy to answer. For example, when a physician selects diabetes, he or she is prompted to indicate mild, moderate or severe diabetes to create an order set. This has been helpful, Landa says, because paper orders meant clinicians could scribble illegible information or simply not answer. The system automatically highlights preferred medications, which has driven use of those preferred medications. Meanwhile, the order set can add documentation. "Whatever the clinician decides about a patient's care, the system records it. That means expanded documentation and the capability to record relevant data."
Landa says there's often a conflict with IT systems: Systems that are easy to use often are very hard to report out of but a structured, data-driven system often is challenging to use. "We're trying to take down those barriers. We're trying to not make it more difficult than it needs to be." With the aging population and the shortage of physicians, "if EHRs slow people down, we're not helping."
The order sets helped AHS in Winter Park, Fla., roll CPOE out to 26 hospitals over a 27-month period, says Philip A. Smith, MD, Adventist's CMIO and vice president of information services. AHS conducted rapid implementation of CPOE, "so we could focus our attention on making incremental improvements," says Smith. Rather than tweak the system during implementation, they decided to get CPOE in all the hospitals and then work to make it better. That plan required very strong executive buy-in, he reports.
While many organizations focus on getting buy-in from physicians, "we really feel it's the local executive that needs to drive the implementation," Smith says. "We have a handful of doctors at every hospital that don't want to use CPOE. We have demonstrated the patient safety benefits of CPOE, however, so it is the responsibility of the physicians to comply. Our goal is better patient care and we've seen some pretty dramatic results."
For example, adverse drug events decreased by as much as 67 percent during one quarter in 2009 over the same period in 2008. AHS has experienced a 95 percent reduction in medication orders that require a pharmacy intervention and 87 percent of medication orders are entered into charts directly by providers. CPOE has led to lower laboratory costs by eliminating redundant and unnecessary tests, and a reduction in length of stay for several conditions, including heart failure.
Everyone is so focused right now on system implementation that it's easy to lose sight of the bigger picture, says Rose. That can "lead down a road of very poor implementations. Facilities might have met the meaningful use criteria but now have huge problems with workflow and acceptance." The key to successful implementation, he says, is to have the end-user understand the end goal of the new system. "Help users understand why this matters to them through the benefits to the provider and the patient."
If done correctly, CDS fosters better workflow, Rose says, which is the key to adoption. One of his mantras is that "everything begins with physician ordering. If you can influence the quality of that order, you can influence the care across the continuum of care. That's why order sets are so important."
Since the majority of costs are driven by what clinicians order, that's the place to improve quality and begin to manage variability in costs, supplies and the expense of reporting, says Rose, adding that this is going to be an increasingly important issue with the new regulations.
The Care Collaborative worked with Zynx Health to create its order sets library. Adventist had evidence-based medicine committees for more than 16 years that only made minimal improvements, Smith says. But, he and his colleagues were impressed with the web-based tools that Zynx developed. The vendor's content had some gaps so the Collaborative developed physician teams to create more extensive content. Now, Rose says Zynx maintains the content by going out to the literature on a "constant, consistent basis" to keep the information up to date. "Few institutions have the staff to continually scour 500,000 MedLine articles a year and then translate the theory into practice."
On the specialty care side, the Center collaborated with Eric Newman, MD, to develop a rheumatology tool called Rheum-PACER (PAtient-Centric Electronic Redesign). PACER collects information from patients using a touchscreen questionnaire and displays information from multiple sources (patient, nurse, physician, EHR) into a series of actionable views used by the nurse and physician.
The motivation for PACER came from Newman, head of Geisinger's rheumatology practice, who identified two key problems: relying on conversation and other unreliable measures to assess a patient's level of pain and function and time constraints to looking at all relevant data. Rheumatologists were trying to make judgments of how treatments were working, but didn't have all the information they needed. And, the physicians said they needed 15 to 17 minutes to review a patient's record but only had two to three minutes.
PACER walks the patient through a questionnaire that takes about eight minutes to complete and then presents all the relevant data in a tabular visual display. As users add to the data, PACER processes the information on the back end. The physician and patient can review trends in outcomes that are meaningful to the patient in relation to temporal displays of treatments that have been used. "There's a very meaningful discussion about how the patient is doing. When the encounter is over, an auto-generated progress note is edited by the physician and then pushed into the EMR. And, a highly tailored after-visit summary is auto-generated as a map for the patient to know next steps and goals.
"Not surprisingly, it turns out that every specialty feels they face the same problem as the rheumatologist," Stewart says. "We have been exploring the use of such tools for primary care." However, primary care is a lot more complicated. "Visits in specialty practices are more homogenous and predictable but primary care represents a form of organized chaos." Stewart has been working to develop similar web applications to assist with conditions common to primary care, such as back pain.
The long-term goal is portable tools that can be deployed anywhere and interfaced with any EHR platform, but Stewart also wants to prove that web applications and patient questionnaires are improving outcomes and patient satisfaction. He's now working on recording physician-patient encounters and, with both parties' approval, digital recordings will be evaluated by experts at Johns Hopkins University in Baltimore.
"We believe that the quality of the conversation really is the mediator of change," Stewart says. "We'd like to know if the web tools we develop really lead to more shared decision making. Our intention is to create value on both the clinical side and the business side."
"The biggest challenges are establishing executive oversight and physician interest," says Eric Widen, MHA, former director of performance improvement, who created the clinical resource management (CRM) program. "Improvement initiatives that are not connected to organizational strategy and a top priority of the executive team typically flounder." At LPCH, the chief operating officer and chief of staff volunteered as sponsors, bringing their authority to the effort.
"To further gain physician interest," Widen says, "we were able to develop unique data views of clinical utilization and variation within defined patient populations to share with them. As most physicians are data driven, they were intellectually curious about the data, leading to brainstorming ideas that generated improvement."
LPCH took a somewhat unusual approach to those clinical improvements. "We started not with which rule or intervention but with the outcomes we wanted to affect," says Longhurst. Some two dozen interventions were put in place.
Not all had the results hypothesized beforehand but many had a measurable effect. In fact, LPCH budgeted $2.5 million in savings in 2010 and $1.7 million in 2011. EMR implementation costs are difficult to recoup, says Longhurst, but financial analysis is starting to create awareness and buy in at the leadership level that interventions have an impact on the bottom line by reducing overutilization.
LPCH has experienced impressive results, such as a reduction in the average number of lab tests per pediatric ICU patient per day of 58 percent by eliminating standing orders. The blood transfusion guideline rule reduced blood transfusions per patient day by 44 percent, or 175 fewer transfusions a year. Changes in the processes for kidney transplants reduced ultrasound exams by 44 percent and chemistry test utilization by 48 percent and saved $700,000 per year.
A major challenge for CDS efforts is alert fatigue—so many warnings that users eventually start to disregard all of them. "It's a real phenomenon and it's important to keep a close eye on it," says Longhurst. One way to avoid alert fatigue is making sure that your alerts are relevant to your patient population.
For example, in pediatrics, most medication errors are due to dosing issues. That's not the case in adult medicine, which sees more drug-drug interaction problems. "We took a conscientious approach when we implemented CPOE to minimize alerts." LPCH turned off drug-drug interaction alerts. Longhurst is comfortable with that decision, even though the facility suffered through its hospital ranking by The Leapfrog Group for doing so. "We have few alerts but when they appear, they're meaningful. We're really trying to ramp up the alerts that provide value to the prescriber."
Longhurst advises others to report back to leadership on metrics that have the most meaning to clinicians. Also, plan ahead "to identify proactive opportunities to improve the system, improve workflow and improve patient care outcomes," he says. "When your organization decides to transform itself by putting in an EMR, you are making an ongoing commitment to increase the optimization of the hospital."
"Clinical decision support (CDS) is an area in its relative infancy in the electronic world and definitely a vital area to focus on going forward," says Jeffrey S. Rose, MD, vice president of clinical excellence informatics of Ascension Health in St. Louis. "CDS hasn't been [implemented or utilized] thoroughly or very well in most places yet. The hope is that after the first phase of meaningful use, people will turn their attention to the whole idea of making improvements at the point of care."
In the post-implementation world, the CMIO's role will be to continue to show the value of these systems, according to Christopher Longhurst, MD, CMIO of Lucile Packard Children's Hospital (LPCH) at Stanford University in Palo Alto, Calif.
But will the HITECH Act and the requirements it mandates really improve patient care? Jason Hwang, MD, MBA, executive director of healthcare at Innosight Institute, a nonprofit think tank in Mountain View, Calif., and co-author of The Innovator's Prescription: A Disruptive Solution for Health Care, isn't so sure.
The HITECH Act and subsequent meaningful use requirements are well-intentioned, Hwang says, but simply implementing computerized physician order entry (CPOE) and EHRs is not enough. "The real benefits of health IT are going to result from new capabilities like data analytics, enabling team-based care, data portability and interoperability, as well as the incorporation of many new streams of data via wireless devices and sensors. Whereas CPOE is merely a modernization of what we do already, these other benefits are truly novel and will bring far more impactful changes to how we can deliver affordable, quality care."
Many facilities have expanded their vision far beyond implementation and developed ways to leverage their health IT for tangible improvements in efficiency and patient outcomes. After all, as Hwang says, "many patients are already searching online for answers to their medical questions before they ever see their doctor. The real challenge will be in convincing healthcare professionals to adopt health IT with the same fervor."
Time will tell as we move toward the next stages of meaningful use whether these initiatives translate to meaningful improvements. CMIO talked to Rose, Longhurst and other leaders at facilities on the leading edge of health IT innovation.
Get smart(er)
Many clinicians have great ideas on how to improve efficiency and patient care but the data needed to implement those ideas weren't accessible. Organizations can design better systems by knowing what ideas need support, says Howard M. Landa, MD, CMIO at Alameda County Medical Center, Oakland, Calif. "We can provide tools in the workflow to both aggregate the data we want and make it easier for people to do the right thing. We can see how we're doing and what's working by diving into the data. This was one of the government's intent with meaningful use."Landa recommends keeping in touch with the "idea people" during the planning phases of CPOE and electronic records. "By including them early in the discussion, we can help them with those ideas without a tremendous amount of work added to our plates."
Alameda County will undergo EHR implementation over the next year. To prepare, the organization began using a relatively simple text-based order entry system that allowed for electronic creation of order sets. Before, orders were simply printed on paper. "We still create paper order sets but we added the decision support you can get from an electronic process and removed order transmittal."
These "smart orders" require answers to certain questions before users can print. The questions are relatively intuitive and easy to answer. For example, when a physician selects diabetes, he or she is prompted to indicate mild, moderate or severe diabetes to create an order set. This has been helpful, Landa says, because paper orders meant clinicians could scribble illegible information or simply not answer. The system automatically highlights preferred medications, which has driven use of those preferred medications. Meanwhile, the order set can add documentation. "Whatever the clinician decides about a patient's care, the system records it. That means expanded documentation and the capability to record relevant data."
Landa says there's often a conflict with IT systems: Systems that are easy to use often are very hard to report out of but a structured, data-driven system often is challenging to use. "We're trying to take down those barriers. We're trying to not make it more difficult than it needs to be." With the aging population and the shortage of physicians, "if EHRs slow people down, we're not helping."
Start with the orders
To drive its CPOE initiative, Ascension Health, Adventist Health System (AHS) and Catholic Healthcare West joined to create the Care Collaborative, a library of more than 1,000 order sets developed at the corporate level and adopted by their physicians. The three organizations, all primarily comprised of community hospitals, pooled their resources to create the order sets.The order sets helped AHS in Winter Park, Fla., roll CPOE out to 26 hospitals over a 27-month period, says Philip A. Smith, MD, Adventist's CMIO and vice president of information services. AHS conducted rapid implementation of CPOE, "so we could focus our attention on making incremental improvements," says Smith. Rather than tweak the system during implementation, they decided to get CPOE in all the hospitals and then work to make it better. That plan required very strong executive buy-in, he reports.
While many organizations focus on getting buy-in from physicians, "we really feel it's the local executive that needs to drive the implementation," Smith says. "We have a handful of doctors at every hospital that don't want to use CPOE. We have demonstrated the patient safety benefits of CPOE, however, so it is the responsibility of the physicians to comply. Our goal is better patient care and we've seen some pretty dramatic results."
For example, adverse drug events decreased by as much as 67 percent during one quarter in 2009 over the same period in 2008. AHS has experienced a 95 percent reduction in medication orders that require a pharmacy intervention and 87 percent of medication orders are entered into charts directly by providers. CPOE has led to lower laboratory costs by eliminating redundant and unnecessary tests, and a reduction in length of stay for several conditions, including heart failure.
Everyone is so focused right now on system implementation that it's easy to lose sight of the bigger picture, says Rose. That can "lead down a road of very poor implementations. Facilities might have met the meaningful use criteria but now have huge problems with workflow and acceptance." The key to successful implementation, he says, is to have the end-user understand the end goal of the new system. "Help users understand why this matters to them through the benefits to the provider and the patient."
If done correctly, CDS fosters better workflow, Rose says, which is the key to adoption. One of his mantras is that "everything begins with physician ordering. If you can influence the quality of that order, you can influence the care across the continuum of care. That's why order sets are so important."
Since the majority of costs are driven by what clinicians order, that's the place to improve quality and begin to manage variability in costs, supplies and the expense of reporting, says Rose, adding that this is going to be an increasingly important issue with the new regulations.
The Care Collaborative worked with Zynx Health to create its order sets library. Adventist had evidence-based medicine committees for more than 16 years that only made minimal improvements, Smith says. But, he and his colleagues were impressed with the web-based tools that Zynx developed. The vendor's content had some gaps so the Collaborative developed physician teams to create more extensive content. Now, Rose says Zynx maintains the content by going out to the literature on a "constant, consistent basis" to keep the information up to date. "Few institutions have the staff to continually scour 500,000 MedLine articles a year and then translate the theory into practice."
Portable, proven tools
"As we continue to innovate, we recognize more and more shortcomings in EHR capabilities," says Walter Stewart, PhD, MPH, associate chief research officer for Geisinger Health System based in Danville, Pa., and director of Geisinger's Center for Health Research. "We also have learned to address these limitations by marrying web applications with the EHR."On the specialty care side, the Center collaborated with Eric Newman, MD, to develop a rheumatology tool called Rheum-PACER (PAtient-Centric Electronic Redesign). PACER collects information from patients using a touchscreen questionnaire and displays information from multiple sources (patient, nurse, physician, EHR) into a series of actionable views used by the nurse and physician.
The motivation for PACER came from Newman, head of Geisinger's rheumatology practice, who identified two key problems: relying on conversation and other unreliable measures to assess a patient's level of pain and function and time constraints to looking at all relevant data. Rheumatologists were trying to make judgments of how treatments were working, but didn't have all the information they needed. And, the physicians said they needed 15 to 17 minutes to review a patient's record but only had two to three minutes.
PACER walks the patient through a questionnaire that takes about eight minutes to complete and then presents all the relevant data in a tabular visual display. As users add to the data, PACER processes the information on the back end. The physician and patient can review trends in outcomes that are meaningful to the patient in relation to temporal displays of treatments that have been used. "There's a very meaningful discussion about how the patient is doing. When the encounter is over, an auto-generated progress note is edited by the physician and then pushed into the EMR. And, a highly tailored after-visit summary is auto-generated as a map for the patient to know next steps and goals.
"Not surprisingly, it turns out that every specialty feels they face the same problem as the rheumatologist," Stewart says. "We have been exploring the use of such tools for primary care." However, primary care is a lot more complicated. "Visits in specialty practices are more homogenous and predictable but primary care represents a form of organized chaos." Stewart has been working to develop similar web applications to assist with conditions common to primary care, such as back pain.
The long-term goal is portable tools that can be deployed anywhere and interfaced with any EHR platform, but Stewart also wants to prove that web applications and patient questionnaires are improving outcomes and patient satisfaction. He's now working on recording physician-patient encounters and, with both parties' approval, digital recordings will be evaluated by experts at Johns Hopkins University in Baltimore.
"We believe that the quality of the conversation really is the mediator of change," Stewart says. "We'd like to know if the web tools we develop really lead to more shared decision making. Our intention is to create value on both the clinical side and the business side."
Plan for ongoing commitment
Lucile Packard's CPOE system uses CDS to help decrease utilization and improve the quality of the care provided, says Longhurst. Those efforts led to LPCH being the first hospital to publish a corollary between a decrease in hospital-wide mortality and implementation of CPOE. "That really reflects the success of our implementation.""The biggest challenges are establishing executive oversight and physician interest," says Eric Widen, MHA, former director of performance improvement, who created the clinical resource management (CRM) program. "Improvement initiatives that are not connected to organizational strategy and a top priority of the executive team typically flounder." At LPCH, the chief operating officer and chief of staff volunteered as sponsors, bringing their authority to the effort.
"To further gain physician interest," Widen says, "we were able to develop unique data views of clinical utilization and variation within defined patient populations to share with them. As most physicians are data driven, they were intellectually curious about the data, leading to brainstorming ideas that generated improvement."
LPCH took a somewhat unusual approach to those clinical improvements. "We started not with which rule or intervention but with the outcomes we wanted to affect," says Longhurst. Some two dozen interventions were put in place.
Not all had the results hypothesized beforehand but many had a measurable effect. In fact, LPCH budgeted $2.5 million in savings in 2010 and $1.7 million in 2011. EMR implementation costs are difficult to recoup, says Longhurst, but financial analysis is starting to create awareness and buy in at the leadership level that interventions have an impact on the bottom line by reducing overutilization.
LPCH has experienced impressive results, such as a reduction in the average number of lab tests per pediatric ICU patient per day of 58 percent by eliminating standing orders. The blood transfusion guideline rule reduced blood transfusions per patient day by 44 percent, or 175 fewer transfusions a year. Changes in the processes for kidney transplants reduced ultrasound exams by 44 percent and chemistry test utilization by 48 percent and saved $700,000 per year.
A major challenge for CDS efforts is alert fatigue—so many warnings that users eventually start to disregard all of them. "It's a real phenomenon and it's important to keep a close eye on it," says Longhurst. One way to avoid alert fatigue is making sure that your alerts are relevant to your patient population.
For example, in pediatrics, most medication errors are due to dosing issues. That's not the case in adult medicine, which sees more drug-drug interaction problems. "We took a conscientious approach when we implemented CPOE to minimize alerts." LPCH turned off drug-drug interaction alerts. Longhurst is comfortable with that decision, even though the facility suffered through its hospital ranking by The Leapfrog Group for doing so. "We have few alerts but when they appear, they're meaningful. We're really trying to ramp up the alerts that provide value to the prescriber."
Longhurst advises others to report back to leadership on metrics that have the most meaning to clinicians. Also, plan ahead "to identify proactive opportunities to improve the system, improve workflow and improve patient care outcomes," he says. "When your organization decides to transform itself by putting in an EMR, you are making an ongoing commitment to increase the optimization of the hospital."
Beth Walsh,
Editor
Editor Beth earned a bachelor’s degree in journalism and master’s in health communication. She has worked in hospital, academic and publishing settings over the past 20 years. Beth joined TriMed in 2005, as editor of CMIO and Clinical Innovation + Technology. When not covering all things related to health IT, she spends time with her husband and three children.