Implementing EMRs: How to Push Physician Productivity

With $19 billion of federal stimulus money in play to modernize healthcare systems, and a mandate for U.S. providers to use electronic medical records by 2013, many hospitals are starting to implement EMRs as a step toward electronic health records.

One message CMIOs should take to heart: “You’re going to lose productivity, there’s no question,” says Howard Landa, MD, CMIO of the Hawaii Permanente Medical Group. However, “with optimization and change management, you can get some of it back.”

Blocking time

Physician productivity took a hit of approximately six weeks when Hawaii Permanente adopted its Epic EHR suite, including an EMR, says Landa. The Honolulu-based network includes a 300-bed hospital, three outpatient surgery centers, and 20 medical office buildings across three islands. There are 650 FTEs in the network, he says.

When Hawaii Permanente took its Epic outpatient EMR live 2004, physicians took a one to two-month productivity hit, Landa says. “We reduced to 50 percent schedule for first two weeks, 75 percent the third week, and 90 percent for fourth week,” says Landa. “By the end of six weeks, we expected them to be back to pre-implementation levels.”

The health system began a gradual EHR rollout on the inpatient side, starting with practice management in 2005. Pharmacy and admission/discharge/transfer went live in 2006, followed by the emergency department and operating room scheduling in 2007. Inpatient clinical documentation and order entry went live in October 2009, Landa says.

Two years after that ambulatory system was first implemented, “people still said they were spending more time with [the electronic forms] than with a piece of paper,” says Landa, especially in primary care. Hawaii Permanente discovered that collectively, physicians were not seeing as many patients per day on average as they did before the Epic system went live. However, Hawaii Permanente also found that, for the first time, physicians were documenting telephone visits and email visits in the EMR, and this may obviate office visits, he adds.

“Between email and telephone visits, I think there’s a lot of gain in productivity that people don’t look at,” Landa says. “The problem is, everyone remembers what takes them longer, nobody remembers what speeds them up. There are very real, tangible improvements in efficiency that we need to call out.”

Mixed results

A September 2009 CompTIA survey of 300 healthcare providers, titled “Healthcare IT Market: Insights and Opportunities,” revealed that 83 percent of respondents using EMRs cited saving time/improving efficiency as a major factor in their decision to adopt the technology. But even with extensive planning and support, those savings can take a while to materialize.

Early technical support was a big part of the EMR rollout at Long Beach Memorial Medical Center, in Long Beach, Calif., says Gary Moreau, medical director of the Emergency Department at the hospital. His hospital is part of the Memorial Health System, a network of five acute-care hospitals in Orange and Los Angeles counties.

Memorial Health System sought physician input early in the EMR selection process, says Moreau. After the system chose Epic, planners provided extensive training prior to launch, which was a phased rollout by hospital beginning in 2007. During deployment at Long Beach Medical Center in 2008, 24/7 IT support was provided throughout the facility. The ED, which includes a trauma center and teaching ER, was upstaffed with three physicians during the first week of deployment, he says.

In spite of this, productivity fell significantly for two to three months in Long Beach Memorial Medical Center’s ER, which comprises 35 physicians and 13 to 14 residents. However, Moreau notes that results have varied across departments in the hospital, and a sister facility where the Epic EMR was implemented in 2007 has reportedly gotten productivity back to pre-implementation levels, according to Moreau.

Built for the future

If an electronic record is to be an effective healthcare tool and not just a prop for regulatory compliance, it can’t just “digitize existing workflows,” says Peter Basch, MD, FACP, medical director, ambulatory EHR and health IT policy at MedStar Health, a network of nine hospitals and affiliated businesses in the Maryland-Washington, D.C., area. “If you just automate current processes, you will never realize the full potential of the technology. But learning these new workflows and processes further reduces productivity at go-live.”

MedStar Health first embarked on its initiative 12 years ago, Basch says. MedStar used internal resources and clinical content consultants to develop its content and clinical decision support framework for its GE Centricity system. The implementation started slowly, but the initiative became an enterprisewide push two years ago, and was deployed by specialty, starting with primary care.

The electronic system has now been deployed in about two-thirds of the network’s ambulatory facilities, says Basch, and is used by some 1,200 doctors in 70 specialties. About four months prior to go-live, Basch and his team work with key staff to develop and optimize forms and practices.

EHR evolution

Robert Dowling, MD, medical director, Urology Associates of North Texas (UANT), in Forth Worth, has gone through two EHR deployments. In 2002, UANT’s first electronic record—built on Amicore’s Penchart EMR product—went live to 30 physicians in the network. It took about three months for UANT physicians to regain their pre-EMR level of productivity, Dowling says.

UANT measures productivity as the number of patients seen, but this is a collective number: “some individual physicians never regained their productivity levels, but some were more proficient,” Dowling says, and others eventually did return to pre-implementation levels.

Eventually, Amicore was bought by Misys, which phased out Penchart. UANT selected Allscripts as a replacement, and in 2008 began a rollout to 50 physicians at 21 sites in four waves. “It was a big bang in physician waves,” he says.

Although UANT worked hard to develop a physician-centric approach to the Allscripts phase-in, according to Dowling, it took longer—about four months—for UANT to regain prior productivity levels. First, Dowling cites the product’s greater complexity and physicians’ “muscle memory:” Those familiar with the older system had challenges adjusting to Allscripts’ more feature-rich interface. Second, intragroup referrals are a large part of UANT practice, and with different physicians in the group adopting the Allscripts system at different times, “patients could see one physician one day who was using the PenChart system, and the next day see a second physician who was using the Allscripts system,” says Dowling. This meant patient information might get lost in the process. The physician-centric deployment sought to minimize this patient safety issue. UANT worked with a third-party integrator to create gap conversions, lowering the risk that patient data would get lost in the transition, he says.

By February 2009, UANT’s entire physician group was using the EHR. After a bumpy start, says Dowling, UANT is looking to continue to interoperate with radiologists and hospitals, and to interface with PACS.

Take your time

Lehigh Valley Health Network, an Allentown, Penn.-based multihospital network comprising roughly 950 beds and 1,200 physicians (400 employed physicians), minimized the physician productivity impact associated with IT implementations at its inpatient facilities by phasing in GE Healthcare EMR modules over the course of several years. Motivated by a desire to boost patient safety and reduce medication errors, the network started with CPOE, rolling it out to the first units in 2001, says Don.

Levick, M.D., MBA, medical director Clinical Informatics, Lehigh Valley Health Network.

CPOE implementation was completed in 2005, and mandated use began in 2006. Electronic medication administration and bar coding modules also were implemented in a staged process, says Levick.

According to Levick, it took some time for physicians to learn to use a computer to look up data such as vital signs and medication administration, which previously had been on paper, and to input orders electronically. To soften the blow to productivity, the Lehigh Valley system phased in the EMR unit by unit, and initially, compliance was not mandatory, says Levick.

“We went service line by service line so we could focus our development of orders and training and support by specialty,” says Levick. “Our physicians have patients on units throughout the hospital. As we brought each unit live, we therefore only impacted the physicians’ workflow for a fraction of their patients.”

“As the physicians became more facile with the system, the workflow impact was lessened when additional units were brought live.”

On the ambulatory side, Lehigh Valley is in the process of implementing an EMR to all owned physician practices and other offices affiliated with the network. To date, they have 200-plus providers online in more than 30 locations.

The EMR is expected to be deployed to all physicians by the end of 2011, says Lori Yackanicz, director, clinical applications, Lehigh Valley Health Network.

The network is deploying the EMR one practice at a time, and works diligently to offset the productivity hit to the bottom line for each office, says Yackanicz. “It’s not just the physician taking the hit, it’s the nurse and clerical staff,” she says. “It can be an enormous cultural change for the practice as a whole.”

Yackanicz says Lehigh Valley’s use of technology has helped bring outpatient practices on board. Both the ambulatory and inpatient facilities use GE systems, and much work has gone into moving data between clinical systems. That interoperability across the continuum of care is big advantage to the provider, she says.

Productivity hits in general are unavoidable, says Yackanicz. However, “if you can play off other things, like the patient portal or interoperability,” she says, “you can get some of that [productivity hit] back.”

The Pros & Cons of Using Scribes: Can They Reduce EMR Productivity Hits?

Practices that employ scribes—and some that don’t—say a well-trained scribe can input data quickly and accurately into the EMR, freeing the physician to spend more time with patients and ensuring that data are accurate. However, still others say the time savings doesn’t offset the added expenses of training and then paying a scribe.   Robert Dowling, MD, medical director at Urology Associates of North Texas (UANT), is a proponent of using scribes. “The emergency room at our hospital has a very successful scribe program,” Dowling says. Scribes are often post-grads or pre-med students who “are highly motivated, can spell and are there because they want to be exposed to the medical environment. They are more educated and motivated than the typical [medical assistant],” he says. “We have tried to deploy this in our practice to improve productivity,” he says. “In our practice, the scribe is logged on as the physician [they share a laptop]. The scribe is at the physician’s hip, listens to the history being elicited and documents in real time. “The physician dictates the exam to the scribe, who enters it on structured note forms. If it is cross-gender, the scribe stands outside the exam room and listens to the dictation. The scribe returns and listens to the counseling, documenting using forms,” Dowling says. “The scribe enters the orders, including meds. The scribe uses the E&M engine to code, and populate the charge interface. The scribe then hands the tablet to the physician, who does his or her own embellishment, usually a synopsis and/or communication to the referring provider.” There are caveats, says Dowling: “Using a scribe requires the physician relinquishing more control. Some physicians have low tolerance for scribe phrases and shorthand.” And part-timers need not apply, says Dowling, because the work requires a high level of system knowledge that differs from facility to facility. UANT makes these trade-offs because  “scribes can help standardize workflows and educate providers. Scribes are much faster documenters than average providers in our group,” he says. A 2007 study by Emergency Medicine Scribes Services, a third-party ER scribe vendor in Santa Barbara, Calif., bears this out, citing ER physician productivity increases of 15 percent to 35 percent, as measured in relative value units (RVUs) per hour, depending on the EMR system in use. Howard Landa, MD, CMIO of Hawaii Kaiser Group, sees both sides of the argument. Hawaii Kaiser does not use scribes because of cost concerns. “The scribe tends to be at your elbow, listening and adding some intelligence to what they’re typing,” says Landa. “and they’re typing directly into the [electronic medical record].” In addition, a skilled scribe “can help the clinician address some of the decision support issues that you can’t with dictation.” Return on investment depends on the type of practice, Landa adds. “If you’re billing per patient, and you are trying to get the highest code that is appropriate for the work you’re doing, having a scribe put in more [information] than you would put in as a physician, because you want to get to the next patient, would be beneficial both clinically and financially in the long run.” At Long Beach Memorial Medical Center, “We’ve talked about [using] scribes from time to time to increase productivity,” says Gary Moreau, Medical Director of the Emergency Department at Long Beach Medical Center. However, the hospital doesn’t use them for cost reasons, Moreau says. Beyond paying another full-time salary, there is a high turnover rate, because often scribes are medical students who might be doing the work as a bridge to the rest of their career, he says. David Stumpf, MD, CEO of EMSS, acknowledges a turnover rate of 25 percent to 30 percent in his business, but says that hasn’t crimped demand for scribes in the ER. In fact, EMSS saw its business triple in 2009, to 15 contracts with hospitals in five states, and Stumpf says he expects another three-fold increase in 2010.