Clinical research studies at Montefiore Medical Center used to take months, required a committee of physicians and data experts, and involved an elaborate research review process. Now, many of these studies are performed by a single clinician in a matter of minutes using “deidentified” patient electronic medical records (EMRs) and software called Clinical Looking Glass (CLG).
“This … software taps into the large pool of EMR data collected here over the past decade,” says Steven Safyer, MD, president and CEO of Montefiore. “While EMRs have the ability to improve quality and reduce costs for single patients, CLG interprets this data for entire patient populations so that we can rapidly check the collective effectiveness of patient safety measures, conduct clinical research, and even comply with federal regulations. We believe it offers a glimpse into how healthcare informatics is shaping the future of medicine.”
2,800 Queries Each Month: Improving Patient Care, Protecting Patient Privacy
More than 700 physicians at Montefiore have taken a three-hour training session to learn how to apply CLG, which has been used to measure the impact of Medicare regulations on rehabilitation patients, quantify the reduction in radiation exposure for emergency department (ED) patients, provide data for professional articles on embolisms and hospitalists, and check on a public health threat.
With clinicians conducting 2,800 CLG inquiries, or analytics, every month, CLG has become an integral part of the culture of healthcare delivery at Montefiore, the teaching hospital for Albert Einstein College of Medicine. Because all of the analyses can be run without identifying the names of the patients, exploratory questions can be undertaken while protecting privacy.
“The goal of these analytics is to gather information and conduct studies that lead to better clinical decision making,” says Eran Bellin, MD, vice president for clinical IT research and development at Montefiore, who was instrumental in designing CLG. “The queries have provided the quantitative evidence for dozens of peer-review journal articles, presentations at professional meetings, institutionwide patient quality improvement initiatives, and programs that benefit entire populations.”
CLG, which was developed at Montefiore, is also being used by the New York City Department of Health in research studies and is being considered for use by the healthcare system of the U.S. Department of Defense.
CLG can slice and dice data sets in multiple ways—immediately. For example, after a clinician generates a hypothesis about a specific disease, CLG can collect data in minutes on patients with that disease who were treated with a specific medicine over a specific period of time and live in a specific neighborhood and then compare this data with patients from a different time period or a different neighborhood or who were given a different medicine for the same disease.
Below are case studies in which CLG has been used as a tool to provide input into public policy debates, develop internal practice protocols, publish peer-reviewed articles, and help respond to a public health warning:
Quantifying the Reduction of Radiation Exposure for ED Patients
To reduce unnecessary radiation exposure to patients admitted to the ED with a suspected pulmonary embolism, a group of radiology and nuclear medicine physicians conducted educational seminars for ED physicians. They showed the ED physicians that for certain embolism patients, high-radiation CT pulmonary angiography was unnecessary, and they could substitute a lower radiation exam, called a ventilation-perfusion scan.
Using CLG, they showed that the number of CT pulmonary angiography scans performed decreased from 1,473 in 2006 (before the educational seminars) to 920 in 2007 (postseminars), for an average reduction in radiation exposure of 23% for each patient.
Publishing Professional Articles on Topics From Embolisms to Hospitalists
Using CLG, colleagues in the department of medicine and epidemiology at Montefiore have shown that tiny filters now commonly inserted in veins to prevent coagulation in embolism patients offer no advantage as adjunct therapy for patients already taking anticoagulation drugs. This internal study became a national medical journal article.
CLG also helped Montefiore researchers demonstrate that the use of hospitalists substantially reduces the lengths of stay for patients with severe illnesses such as stroke and pneumonia while maintaining quality care and quality resident education. This study was published in the Archives of Internal Medicine.
Measuring Medicare: A Policy Case Study
In 2005, the Centers for Medicare & Medicaid Services announced a new rule for rehabilitation hospitals, requiring that a large percentage of their patients must have one of 13 specific diagnoses or the facilities could lose reimbursement. Concerned about the rule’s potential impact on patients, a group of physicians at Montefiore, using CLG, performed a retrospective study of discharges from the hospital during the year before the rules were implemented and discharges the year after the rule was enacted.
They found that, in general, “restricting access to inpatient rehabilitation on the basis of diagnosis alone” resulted in patients being readmitted sooner and dying sooner. Their study called for broader, evidence-based guidelines to allocate rehabilitation services. It was published and augmented the case against the rules.
CLG Used in Heparin Scare
Last year, during the made-in-China heparin scare, the FDA did a random review of hospitals, and Montefiore Medical Center was chosen for inspection. The FDA review revealed that Montefiore had properly notified its physicians and evaluated its stock. However, Montefiore then did a rapid, hospitalwide study to check whether patients using heparin that year vs. the previous year had a longer hospitalization or died sooner. Clinical staff keyed in “heparin” on the computer and tracked all the patients who had been given the blood-thinning drug in December 2007 and January 2008 and compared their length of stay and mortality with patients who had been given heparin in December 2006 and January 2007. The answer: There was no increase in either measure—and no cause for alarm.
Monitoring, Improving Care for 14,000 Diabetes Patients
Using CLG, physicians at Montefiore’s 25 ambulatory care centers have been able to identify 14,000 diabetes patients, track their care and progress over time, and compare improvements in blood glucose and cholesterol levels of the patients at one center vs. another and even one physician vs. another. Comparisons are important because some Montefiore physicians practice outcome-based medicine linked to pay-for-performance incentives—physicians whose patients have lower scores get more pay. The pay-for-performance initiative has greatly helped lower blood glucose and cholesterol levels in these patients. Without CLG, Montefiore would not be able to monitor and compare scores for this many patients.
Patience Results in Intellectual Growth
Not all 2,800 CLG queries each month bear immediate fruit. A clinical hypothesis is posed, a group of patients with a certain condition is identified, CLG runs the data and, for one reason or another, there is a dead end. When this happens, the Montefiore investigator has an intellectual home to go to for collegial advice. The division of general medicine, for example, holds bimonthly lunches at which ideas and experiences are shared, avenues for productive exploration recommended, and new CLG inquires are rigorously critiqued by peers.
“CLG allows all clinicians and hospital administration to test hypotheses that just a few years ago would have been impossible to do. It democratizes the research process. To be able to identify, within minutes, a cohort of patients, qualify them by a specific disease, medication, clinical event, race, time period, or neighborhood provides a new paradigm for population-based medicine and a new level of importance for healthcare informatics,” says Bellin.
— Source: Montefiore Medical Center