When President Barack Obama cites Mayo Clinic as a model healthcare provider, he praises its “smart” practices that offer patients the best possible care at below-normal cost. Mayo’s expertise in treating disease is well-known. But the presidential accolades underscore Mayo’s pioneer work in an emerging science of healthcare delivery.

Ognjen Gajic, M.D., is a leader in that effort as head of a clinical-research laboratory working to improve quality and safety in Mayo’s intensive care units. Using in-depth medical-records research, informatics, epidemiology and systems engineering, Dr. Gajic’s lab, called “METRIC” (Multidisciplinary Epidemiology and Translational Research in Intensive Care), is discovering the best medical practices for ICUs.

“These system interventions may save more lives than many laboratory science breakthroughs,” Dr. Gajic says. “The smart environment can also prevent us from wasting health care dollars.”

The first 24 hours after acute illness or injury are critical. If poor decisions are made during that time, a patient is likelier to experience complications leading to major organ failure. Such complications can prolong hospital stays and result in death or severe loss of function, necessitating long-term care.

Mayo’s ICU protocols have the potential to change the way critical care medicine is practiced. Good decision-making in the ICU is essentially about pattern recognition; small changes in a patient’s condition can form a recognizable pattern forecasting ICU complications. The Gajic lab’s research is uncovering these patterns to facilitate appropriate and timely decision making in emergent situations.

“We are trying to make health care delivery in the ICU a science,” Dr. Gajic says. “Because these are our sickest patients, delivery of health care in the ICU is a high-risk enterprise. Errors of process are common and even small ones can have enormous consequences. We want to facilitate timely resuscitation and avoid healthcare-acquired complications.”

A hospital GPS: nudges to prompt safe decisions

Dr. Gajic’s lab started work in 2005. Emerging discoveries have a potential to substantially improve Mayo’s critical care outcomes and cut costs. Some of METRIC’s suggestions are as simple as requiring doctors to think twice before ordering blood transfusions. Others involve computer simulations in which off-duty doctors practice emergency decision-making, much as airline pilots do. Dr. Gajic likens METRIC’s work to a hospital Global Positioning System: Just as GPS offers relevant and timely information to drivers on busy streets, the Mayo systems guide ICU practitioners to the best choices.

“The system is designed to help the providers,” Dr. Gajic says. “They don’t drive for you.”

METRIC’s cornerstone is continuous and comprehensive monitoring of patient outcomes in Mayo ICUs. This commitment to quality control has led to some simple interventions. Ensuring the safety of critically ill patients admitted during the night required a staffing change. Previously, critical care fellows – doctors who are training to specialize in critical care – provided care in the ICUs at night, calling in specialists if needed. But after introducing a night shift for critical care specialists, Mayo saved an average of $5,000 per patient and almost $15,000 for very sick patients admitted after hours.

“Adding human expertise makes a big difference,” Dr. Gajic says.

Another innovation helps clinicians decide when it is safe to release patients from intensive care to regular hospital wards. METRIC developed a model that predicts a patient’s risk of suffering complications and returning to the ICU. If that risk is 15 percent or higher, ICU staff are directed to discuss the case further. They may decide to give additional instructions about the patient to the receiving service, send the patient to a ward with specialized monitoring, or even keep the patient in the ICU an extra day. As a result, the number of patients re-admitted to intensive care within 24 hours fell from 4 percent to 1 percent.

Dr. Gajic calls such changes “little nudges” that help providers make safer decisions. A similar “nudge” reduced Mayo’s use of blood transfusions. Physicians have routinely ordered transfusions when patients’ hemoglobin (the oxygen-carrying pigment in the red blood cells) fell to a certain level. But clinical studies suggest that low hemoglobin alone doesn’t justify the risks of transfusion. Mayo started requiring doctors considering transfusion to check off a second factor — such as bleeding, low blood pressure and “other” — with no further explanation necessary.

“It cues doctors to think before they order the transfusion,” Dr. Gajic explains.

Within three months, Mayo saw a dramatic decline in transfusions and transfusion-related complications, with no detriment to patient outcomes.

Data mart “sniffers”

Other tools are more complicated. Vitaly Herasevich, M.D., Ph.D., a senior research fellow in the METRIC lab, has developed “sniffers,” sophisticated computer algorithms embedded in software that, like a bloodhound, detects potential problems. Traditional bedside monitors trigger an alarm when a patient’s vital signs change. But the alarm is just noise; it can’t specify the problem. “And 95 percent of these alerts are false alarms,” Dr. Herasevich says.

The “sniffers” software crunches patient data that is fed continually into a METRIC “Data Mart.” When the computer detects a problem, it sends a text message to a clinician. For example, if a patient on a ventilator is receiving too much air (based on the individualized estimation of lung size according to gender, height, type of lung injury and lung stiffness), the computer quickly texts a respiratory technician to check that patient’s ventilator setting. This helps prevent a condition known as ventilator-induced lung injury (VILI), in which patients’ lungs are repeatedly overstretched and injured by large air volumes. Sniffers provide specific, fast alerts — critical for “detecting complications like VILI, severe sepsis and acute kidney injury,” Dr. Herasevich says.

An ICU dashboard to ease information overload

The reams of data streaming through Mayo’s Data Mart highlight another ICU challenge: information overload. Conventional wisdom suggests the more information physicians have, the better. But in critical care, where life-and-death decisions must be quick, “we are overwhelmed by masses of information and we make bad judgments,” says Brian Pickering, M.B., B.Ch, a critical care fellow in the METRIC lab.

A leader in the use of sophisticated electronic medical records, Mayo is taking that technology to the next level and studying what pieces of information doctors actually need to make safe and timely decisions in critical care. After critical-care events, often in the middle of the night, Dr. Pickering surveys bedside providers on which data they used to make treatment decisions. Out of thousands of pieces of data, the doctors relied regularly on about 30. The data’s relative importance varied according to patient demographics and condition. Combining those findings with in-depth medical records research, the METRIC lab created an information-scoring system called “DUM” (Data Utilization in Medical Decision-making) that prioritizes information for ICU syndromes such as sepsis, coma and bleeding.

“We’ve managed to identify which subsets of data are considered valuable in different clinical contexts,” Dr. Pickering says.

The next step is to help ICU clinicians get that customized information quickly. To that end, METRIC is developing a new version of an ICU patient monitor — a “21^st century ICU dashboard” — that will display information relevant to specific patients.

“We expect, over time, to prove that this display facilitates decision-making, reduces cognitive load on providers and ultimately leads to safer care of acutely ill patients” Dr. Gajic says.

Mayo is also using data from real patients to create computer simulations with which off-duty doctors can make virtual treatment decisions. Their decisions will be compared with the patients’ actual treatment and outcomes to give METRIC researchers further insight into the cues physicians need for better decision-making.

A natural leader in the science of health care delivery

It all adds up to a redesign of critical care services, aimed at helping ICU providers make the best decisions. Mayo is uniquely qualified to be a natural leader in the new science of health-care delivery. With 206 intensive-care beds in ten ICUs, Mayo is the nation’s second-largest critical-care provider and has long fostered a team approach and commitment to research that benefits patients. Although health-care delivery has been formally prioritized in the U.S. for only a decade, “The Mayo Brothers started this tradition when they founded Mayo,” Dr. Gajic says. “Mayo has been among the first institutions in the world to adopt health care delivery as a science. We are proud to have the culture at Mayo that allows me to do what I do.”

METRIC’s work has lowered critical-care costs at Mayo but the lab’s primary focus is improving patient care.

“None of what we are doing is containing costs for costs’ sake,” Dr. Gajic says. “The mantra that ‘Patients Come First’ really works here at Mayo.”

He speaks from experience. Three years ago, Dr. Gajic collapsed outside Mayo’s emergency room with anaphylactic shock from multiple bee stings and was close to cardiac arrest.

“In many places I would have died,” Dr. Gajic says. But at Mayo, “I felt safe. And although what I do for a living is to ask questions, I questioned nothing. I think that’s what happens when you are critically ill. You don’t question, you just want to be safe.”

After just one day in a Mayo ICU, Dr. Gajic was back at work. “Being a physician here is a privilege,” he says with a smile. “But being a patient here is a bigger privilege.”

— Barbara Toman, September 2009