Better Clinical Analytics Means Better Clinical Care
Healthcare providers are incorporating BI capabilities in new ways to improve patient outcomes.
Southeast Texas Medical Associates, a midsize doctors' practice in Beaumont, Texas, has slashed its hospital readmission rate by 22% in the last year. It also has reduced the number of visits it gets from diabetes patients around the holidays by 15% (such patients sometimes stop watching their diets and testing their blood around the holidays). The improvements--mostly the result of fewer complications--came about because the practice, known as SETMA, has found a way to keep its patients healthier.
SETMA hasn't stumbled onto some miracle cure. Its doctors are using business intelligence and analytics to fine-tune their care. The practice is employing tools from IBM Cognos to improve therapeutic and preventive care programs and boost staff productivity, in addition to reducing hospital readmissions. Other healthcare providers are following suit, employing everything from data warehouses at major cancer centers to mobile applications used by oncologists and their patients.
SETMA, with 29 primary care physicians in three locations, has used a NextGen EMR system since 1998. Two years ago, the practice invested in the Cognos tools to analyze data extracted from patient records. The Cognos products include dashboard software and tools for trending and auditing data. The extracted data resides in a data warehouse with records for 65,000 SETMA patients, including 7,500 patients with diabetes and 26,000 with hypertension. The data analysis tools are being used not only by BI specialists but by senior management and clinicians.
The tools let SETMA analyze the health and treatment trends of patients and audit the performance of SETMA clinicians to ensure that they provide evidence-based standards and adhere to best practices and quality-of-care measures from a number of sources.
Those quality measures include the Healthcare Effectiveness Data and Information Set from the National Committee for Quality Assurance, the Centers for Medicare and Medicaid's Physician Quality Reporting Initiative, and care benchmarks from the National Quality Forum.
For example, SETMA's use of the Cognos tools for daily audits on quality measures provides clinicians with specific goals for the patients they see during the day. The tools ensure that screenings and tests are up to date, and cover preventive care programs for individuals with chronic health conditions such as diabetes.
The analysis has also helped SETMA identify "interruptions of care" and act quickly to resolve those issues. Care interruptions can range from patients missing tests or follow-up appointments to failing to have their medications or other treatments adjusted to address a change in health.
"Healthcare is behind the times in the use of BI," says SETMA CEO James Holly. "But everything that BI can do in other industries, it's doing for us at Southeast Texas Medical Associates, leveraging data to improve care."
How Tools Help
Using the analytics tools, SETMA clinical support staff can run patient reports prior to a patient's visit to review an individual's health status, determining whether screenings or blood tests are due, so that those tests can be ordered and completed before the visit. With much of that necessary work done prior to the office visit, doctors can discuss the findings with patients and develop a care plan during the scheduled appointment.
That proactive effort even includes sending reminders to diabetic patients around the holiday season to encourage extra vigilance about exercise, glucose testing, and diet, says Holly. One year after implementing these preventive efforts for diabetic patients to better control their blood sugar and related complications, SETMA had a 15% decrease in patient visits during the holiday season.
Data analysis lets SETMA identify issues that can play a role in patient outcomes. Hospital readmission shortly after discharge is one good example of an issue doctors follow closely. According to a study last year by the University of California San Francisco, nearly one in five Medicare patients in the United States has an unplanned readmission to the hospital within 30 days of being discharged.
Within the first six months of using the Cognos tools, hospital readmissions of SETMA patients were reduced by 22%. SETMA used the analytics tools to compare patients who didn't get readmitted to those who did, taking into account patient characteristics such as age, gender, ethnicity, follow-up care, and how soon follow-up care was received once the patient left the hospital.
Using the Cognos tools to analyze patient EMR data, ICD-9 billing codes, and other information, SETMA identified patterns that showed groups of patients who are more likely to be readmitted, and the factors that contribute to that. SETMA saw that patients who live alone and those in lower socioeconomic groups were at greatest risk for readmission.
The data revealed, for instance, that patients who live alone are less likely to adhere to their follow-up care instructions while lower income patients often can't afford medications prescribed for ongoing treatment once they're out of the hospital.
The findings prompted SETMA to institute new post-hospitalization treatment plans that include help setting up immediate at-home care and interventional support services for patients who live alone. SETMA also created a foundation to help low-income patients pay for their medications.
SETMA spent about $500,000 on the Cognos project, says Holly. "It was expensive, but the payoffs are enormous, and we're just scratching the surface," he says.
Before using the Cognos tools, it would take SETMA about 36 hours to run "daily" patient-encounter reports using the NextGen EMR database and query functions. That means SETMA was always at least a half-day behind in seeing if patients had received the evidence-based care they should have gotten from clinicians based on their health situation. Now applying the Cognos tools to data in the data mart, complex analysis can be completed in seconds, says Holly.
Analytics For More Personalized Care
Meanwhile, Moffitt Cancer Center in Tampa, Fla., is among larger healthcare providers focusing on analytics to personalize the treatment of cancer patients. Moffitt recently announced it is collaborating with Oracle on a new health and research informatics system in which analytics plays an important role.
The new system, built with Oracle Health Science technology, will support Moffitt's Total Cancer Care program, a comprehensive approach that provides individualized, evidence-based care. With patients' consent, Moffitt aggregates and analyzes data, including diagnoses, treatments, follow-up care, and bio-specimens such as tumor samples from thousands of individuals cared for at nearly 20 cancer treatment hospitals in the U.S. that participate in the Total Cancer Care program.
For Successfuly Using BI
1. Make sure data being analyzed is properly scrubbed and that data fields are normalized and accurate so that queries reveal valid comparisons (apples to apples, not apples to oranges)
2. Realize that insights gained from data analysis will often lead you to do further inquiries to drill down even more on the findings
3. Give as many users access to the BI tools as possible to stretch the value of the tools. Create a strategy for teaching these users how to run reports to glean insights into possible areas for care improvement
4. Make sure the displays of your analytics are intuitively designed so that they require little explanation
Data about the specimens, treatments, and patient outcomes are analyzed for biological or clinical elements that could advance the care of treatment for other patients. If new discoveries come along, including successful new treatment protocols, doctors can contact other patients about the advancements and the possibility of participating in clinical studies.
Among the Oracle products and technologies being used by Moffitt for its "next-generation health and research informatics system" are Oracle Health Sciences Enterprise Healthcare Analytics. These tools include data warehouse and analytics products for healthcare data that comes from healthcare providers' e-health records and other internal and external clinical applications, as well as back-office, research, and other systems.
Oracle and Moffitt are working on adding new data sources to the Oracle Healthcare Data Warehouse Foundation data model, including DNA sequencing data from patient tissue, including malignant tumors. The new system will also support the analysis of DNA sequencing data and other sources of patient data, in the hopes of providing more rapid insights into the effectiveness of treatments.
The analysis could yield new findings about the effectiveness of specific cancer treatments on patients with common bio-markers and other similarities, enabling researchers and clinicians to more quickly advance personalized treatments for others.
Moffitt is also using the Oracle Healthcare Data Warehouse Foundation data model to normalize and aggregate data so that data quality is optimal for analysis. "You want to compare apples to apples, not apples to oranges," when analyzing patient data, says Kris Joshi, Oracle's VP of healthcare market strategy.
For instance, if two different labs conducting the same test express results in different values (say, milligrams versus micrograms), the data must be standardized so that statistical calculations are accurate. "A doctor could look at both test results and understand they're the same, but once those different values are in aggregate, it's hard to fix," Joshi says. Oracle's platform transforms and normalizes data so that analysis is accurate.
Decision Tool Offers Free Cancer Help
Even healthcare providers that don't have the manpower, IT budgets, or tech know-how to collect and analyze their own patient data can tap the decision support resources of others. CollabRx and the American Society of Clinical Oncology recently released a free, Web-based decision support tool to help cancer patients and oncologists identify the best choices of tests, treatments, and clinical trials based on details about an individual's cancer type and other information.
The new Targeted Therapy Finder--Melanoma app is the fruit of collaboration between ASCO and CollabRx, a California startup launched about three years ago by Jay "Marty" Tenenbaum, a melanoma survivor, tech entrepreneur, and former chief scientist at Commerce One. While the app for melanoma, a type of skin cancer that's often deadly, is the first tool to be offered, the collaboration between CollabRx and ASCO could lead to decision support tools for other cancers as well, says Tenenbaum.
BI In The Real World
What Southeast Texas Medical Associates Achieved
Reduced hospital readmissions by 22%
Cut office visits of diabetic patients during the holidays by 15%
Allowed the practice to publicly report by provider name over 200 quality metrics
Reduced the time to run analysis of daily patient encounters from 36 hours to minutes
The Web tool, which can run on mobile devices, helps doctors and patients identify diagnostic tests, treatments, and clinical trials associated with specific subtypes of melanoma and other characteristics of the patient, says Tenenbaum. It can be used by busy oncology practices that don't have big IT budgets, as well as by patients who don't live near major cancer centers, he says.
Patients or doctors enter information about an individual's melanoma, including the disease's stage, origin, metastatic sites, and genetic mutations, into the app, and it narrows down drugs and clinical trials to consider. The app provides definitions of terms and descriptions of specific mutations, and hyperlinks connect users with additional details about drugs and clinical trials.
The melanoma tool is powered by an expert knowledge database called Cancer Commons, which was set up by CollabRx and runs on Amazon's cloud computing service, says Tenenbaum. Cancer Commons incorporates data from molecular disease models (MDMs), including a melanoma disease model that was recently published at PLoS One, an international, peer-reviewed, open-access research publication site.
The therapy-finder application transforms the contents of the knowledge base into personalized, actionable information that can help patients and physicians make informed decisions about tests and treatments, says Tenenbaum. Each cancer has a decision algorithm, represented in a set of decision/workflow tables, or DWTs.
The DWT technology, developed by CollabRx, "is in essence a set of flowcharts that integrates the semantic contents of the MDMs with the Web-face of each therapy-finder application," he says. The therapy finder itself is a Ruby on Rails program that imports the MDMs and DWTs and combines them into an efficient Web app.
As part of the partnership, ASCO is providing CollabRx with access to its published melanoma content, including data presented in abstract format at the society's annual meetings and study results described in peer-reviewed journals such as the Journal of Clinical Oncology. The vision for the Cancer Commons database is to include outcome information about how patients are responding to various treatments so that insights--for instance, the success of certain drugs with various cancer subtypes--can be disseminated more quickly, says Tenenbaum.
"Every day in oncology, there are thousands of experiments that go on but no one captures that data," or has the ability to analyze it, he says. That includes patients who receive chemotherapy cocktails that target specific subtypes of cancer.
The ability to analyze and share outcome data on patients who have uncommon combinations of genetic mutations, chronic conditions, and other characteristics could make it easier for oncologists to zero in on treatments for patients who have rare similarities.
Clinical trials for new treatments are often considered failures if only a handful of patients out of thousands respond positively to the new drugs, says Tenenbaum. But for the few patients who do respond well, those trials are hardly failures.
When Tenenbaum received his melanoma diagnosis several years ago, it was as though "my hair was on fire," he recalled. The anxiety of receiving a serious health diagnosis is often made worse by the fact that "five different doctors will give you five different suggestions" for treatment.
Lives can be saved by getting patients matched up with the best available drug options sooner. "That's what we're trying to do," says Tenenbaum.