Our QualOptima Health Informatics Platform has evolved from Quture’s extensive consulting engagements. Quture is different from our competitors offering information technology products based on 40+ years of being responsible for identifying solutions to poor patient outcomes. To recommend any treatment or prospective process of care, Quture developed the most sophisticated and robust system of data collection and analytics in health care. Accurate diagnosis including thorough root cause analyses are essential to identifying and evaluating potential solutions.
Quture’s most recent consulting engagement provides extraordinary insight into how we have developed QualOptima to empower Enterprise Value Management (EVM). We were on-site using our original peer review software product reporting to the hospital’s senior management and board of trustees. The hospital’s medical staff had peer reviewed over 500 cases of suboptimal outcomes, consistently finding “care appropriate” without any opportunities to improve. Quture’s QualPRO expert panels of medical specialty physicians identified opportunities to improve in every case, as well as suboptimal clinical performance in over 80% of the identified cases. Quture’s system of clinical indicators further identified five (5) times as many significant cases of suboptimal outcomes for the same period of time from the hospital’s database.
This would have been just another external peer review project if we had not recognized, working with the board of trustees, that our technologies could align quality with the hospital’s strategic plan. We learned how to evolve peer review processes to value-driven analysis. Quture’s EVM solution began when we aligned our research and methods of enterprise management from other industries with factual situations and challenges in health care.
One evening at a monthly board meeting, reporting on the status of our engagement, we learned that this hospital’s board had spent several million dollars on a strategic plan that focused on increasing admissions through their emergency department. The available database demonstrated that neither senior management nor the board, and most notably not even the hospital’s organized medical staff, knew anything about clinical performance or outcomes by their emergency department.
The hospital’s ER was contracted to one emergency specialist physician, who owned a group of contract emergency medicine physicians to staff the hospital emergency department. The group conducted their own quality and outcomes analysis using their own internal processes. As with most hospitals, their contract did not require access to the quality information which was not shared by the outside emergency medicine group. The group refused our request for cases they had identified for peer review.
Quture received authorization from the board, legal counsel and senior management to conduct an external peer review of the emergency department. Based on the strategic plan of the board, we agreed to focus on whether the existing emergency service provided the anticipated level of value to the organization. This engagement became the initial prototype of an EVM engagement and use of our informatics technology.
We coordinated our review with several groups of distinguished and nationally- recognized leaders in emergency medicine based on clinical, operational and financial outcomes. Our standard clinical indicators were supplemented for these additional factors, and we constructed electronic collection, integration and aggregation of a strategic database for analysis. This consulting engagement was an early example of Quture’s development of a “2nd generation database,” now often called a “data lake.”
Quture’s review and analytics methods led to the discovery of profound disquality performed by this emergency medicine group. The extraordinary deficiencies in care and outcomes in all three (3) tiers were reported to the board and senior management. Together, we eventually visited with the ambulance drivers of the metropolitan countywide service. They openly explained that they had independently determined that the emergency care was so poor that they only brought two types of patients to this hospital: those who couldn’t pay or were dead or dying.
Using what is now QualOptima, Quture was able to work with the board and senior management to cancel the emergency department contract without any negative financial impact. Meticulously designed and performed analysis saved the hospital millions of dollars that could have been claimed through litigation for premature contract termination. Quture then coordinated transforming the department to a new physician group with financial incentives tied to their contract and payment. Assuming the client’s strategic plan was correct, the potential increased revenue to our client is extraordinary. And the strategic plan is now measurable!
This was the beginning and now is the future of Enterprise Value Management.
When Quture began expansive redesign and reengineering of our software and introduced QualOptima in 2010, one of the basic enhancements was to aggregate data from disparate databases into a 2nd generation database. We had already learned from peer review engagements the value of capturing existing electronic data from existing software of other vendors in niche products. For example, we began searching electronic surgical databases of various vendors to assess perioperative indicators such as operative and anesthesia times, blood loss, and pathology databases for surgical performance and outcomes.
This commitment to create a 2nd generation database was one of the primary factors in transitioning to the InterSystems platform. We were forced by the realities of what is known as “interoperability” to migrate our technologies to what is now the HealthShare Platform founded on the Cache multidimensional database. The performance of the Ensemble “interface engine” to collect data from disparate databases of multiple vendors in various data formats integrated and aggregated into the Cache database is central to QualOptima technologies. Add the power of analytics and rapid data displays with DeepSee, and it was a no brainer.
These were days long before the impacts of data analytics and the realization of what is now the tsunami of data collection and availability for analytics. What we call 2nd generation databases are now often called data lakes. A “Data Lake” is a repository of data-centered architecture and informatics tools capable of storing and analyzing vast quantities of data in various formats. Data from webserver logs, data bases, social media, and third-party data can be ingested into these Data Lakes.
Compare Quture’s technology during this engagement and what we discussed as the EVM capabilities to collect and aggregate data lakes with the QualOptima Health Informatics Platform. (See the recent Note “Enterprise Value Management: Integrating Value Management with Corporate Strategy.”)
Then compare Quture’s analytics technologies working on this emergency department assessment. These were also days when we were just beginning to include machine learning and other technologies, now recognized in aggregate as artificial intelligence (AI), in QualOptima’s analytics engine. Quture began working and incorporating machine learning in 2008 with a leader in natural language processing (NLP) in Jacksonville, Florida. One of our early decisions, in view of the significance of our technologies and business application strategies, was to recruit Sherif Elfayoumy, Ph.D., as our Chief Technology Officer. Dr. Elfayoumy was one of the early architects of the NLP product Quture was exploring. When Quture committed to the Application Partnership with InterSystems, we were unaware of their purchase of iKnow, the machine learning technology in InterSystems’ HealthShare product. Quture has devoted our product design and strategies to machine learning, integrating the total array of machine learning technologies, now incorporating the iKnow unsupervised learning technology of iKnow.
Quture announced in some detail how our product operates as the QualOptima Health Informatics Platform on July 19th. See http://www.quture.com/qualoptima-health-informatics-platform
Business strategies require positioning QualOptima as the solution to the most urgent and rapidly approaching and inevitable needs of our customer segments. Development of the QualOptima Health Informatics Platform provides the logical progression for customer use of data lakes. The fundamental business strategies depend upon the goal of aggregating all this data into a common data platform. Many entrepreneurial enterprises are now working with data lakes to improve health care by empowering customer segments with data transformed to information. While most of these are operated by information technology experts with little if any health care experience, Quture has evolved from the understood needs of our customers – providers, patients, payers and insurers.
Before launching into real-time applications of data lakes, retrospective opportunities must be fully developed with extensive knowledge derived for use in real-time at point of care. The logical and scientific advances of empowering customer segments from data lakes has three (3) essential progressions to achieve the potential of these data repositories and technologies:
- Retrospective Quality & Outcomes – Retrospective data analysis to examine performance and outcomes in the context of value to develop, implement and monitor optimal clinical processes from the unique but integrated perspectives of quality, risk, and current clinical competence of providers;
- Real-Time Point-of-Care – Empower clinicians at point-of-care with aggregated complete patient data reformatted for rapid insight and comprehension – and preferably involve engaged patients with individual-controlled electronic records and care plans; and
- Real-Time Precision Predictive Care – Integrate medically sound predictive modeling at point-of-care, personalized to the meFactors© of the individual and used to engage the patient in data-driven decisions for their health and care.
In the example discussed above for analysis of the hospital’s strategy centered on the emergency department, we used our system of clinical indicators to identify cases and patterns of suboptimal outcomes. QualOptima now has matured the capability to dramatically use its 2n generation data lake on the common data platform to exponentially advance both the speed and insights from retrospective performance and outcomes analysis. Quture’s Value Data Centers are now empowered to use the total array of QualOptima’s processes:
- Performance & outcomes metrics – medical specialty organization and CMS metrics
- Clinical quality indicators – traditional process to identify cases for peer review
- HFACS error classification – optimal event and near miss process
- Electronic triggers with NLP (natural language processing), – electronic searching for specific defined terms and phrases in narrative text
- Computational trigger algorithms – advanced mathematical and computational tools using triggers as signals
- Unsupervised machine learning – total array of unsupervised and machine learning (a form of artificial intelligence).
- Patient experience – new technologies replace “patient satisfaction” tools with innovative patient experience methods
- Mortality & claims data review – traditional morbidity & mortality quality reviews and risk management causation of claims (medical malpractice)
One way to consider these improved technologies is to think of clinical indicators and triggers as recognized problems. For example, analysis depends on identifying known problems such as medication overdoses. Even here, using NLP QualOptima can search for the drug Narcan used to reverse medication overdoses. Quture has already advanced risk and quality capabilities from searching massive narrative text.
However QualOptima adds the ability to search for what might not be expected and relying on specific clinical indicators or triggers. Unsupervised learning analyzes the 2nd generation data lake for hidden patterns, essentially creating potential categories of clinical problems not initially suspected.
The combination of unsupervised learning and triggers as signals can be used to identify unrecognized problems such as described by Isaac Kohane, MD, PhD, at Harvard. He explains how health care is missing huge signals by not paying attention. He gives the example of two top hospitals in Boston which had an 18% increase in heart attacks. Most dramatically, Dr. Kohane points out that “not only did no one ever notice it, we actually caused it by giving the drug Vioxx to our patients” (causing cardiotoxicity).
Quture embeds our own proprietary algorithms developed over years of reviewing clinical performance and outcomes. We also follow advances reported in the medical literature to solve specific problems, such as the most frequent medical error – failure to diagnose.
Some of the most important work being done using triggers is by Daniel R. Murphy, M.D., and his associates developing trigger algorithms to proactively, efficiently and effectively identify delays in cancer diagnosis. His team has also reported the capability of trigger algorithms to identify failures to follow-up in patients with abnormal radiology studies. Using this algorithm at Baylor College of Medicine and the Michael E. DeBakey Veterans Affairs Medical Center’s Center for Innovations in Quality, the researchers found 131 patients whose criteria made them at higher risk for a delay in care. Chart reviews confirmed that 75 of these patients truly needed action but had not received it within the 30-day period.
As we will discuss in the next Note, this use of data lakes retrospectively provides the ability to then use QualOptima technology at point-of-care in real time. Dr. Murphy’s initial studies were retrospective but are now used as trigger algorithms to identify patients in real time who need care to improve their outcomes. Dr. Murphy explains: “Our trigger algorithms can act as a safety net to detect patients who are at risk for delays and alert designated medical personnel of this risk.”
Another advancement with QualOptima for Value Data Centers derives from using specific metrics along a clinical performance process to analyze outcomes. Value as a formula of quality and cost must evaluate the payment penalties from certain non-reimbursed outcomes, such as Hospital-Acquired Conditions (HAC). This is one of several federal initiatives of the Centers for Medicare and Medical Services (CMS) targeted for significant payment reductions. Quture demonstrated QualOptima’s improvement of clinical, operational and financial outcomes in a joint project with Excelcare, a Quture strategic alliance, aimed at one specific type of HAC, catheter-acquired urinary tract infections (CAUTI) in intensive care at Niagara Falls Memorial Medical Center (NFMMC), Niagara Falls, New York.
The NFMCC project implemented an evidence-based optimal clinical process for prevention of CAUTI in the intensive care unit. Quture added a test and metric in the treatment protocol adopting the Centers for Disease Control and Prevention (CDC) evidence-based guideline and identifying evidence-based metrics implemented in the Excelcare data collection and aggregation system. Adding one key metric of this protocol for bladder screening dramatically improved the optimal clinical process for performance measurement and impact on the goal of zero CAUTI outcomes.
The last example of Quture’s logic for initial use of data lakes retrospectively comes from our Clinical Trial at the University of Miami, Miller School of Medicine and Jackson Memorial Hospital. Using QualOptima analytics, Quture examines post-operative nausea and vomiting (PONV) in the preventative use of anti-emetic drugs. Data from this QualOptima system not only measures clinical performance and outcomes, it provides Quture’s unique system of root cause analysis from electronic data rather than relying on manual chart review. We will discuss in the next Note how this model will be used in the real-time applications of the QualOptima 2nd generation data lake.
PONV and analysis of maintenance of body temperature in perioperative medicine are also examples of how Quture will provide clinical rather than billing data as PQRS measures for MACRA. Reporting quality and cost measures in the future of this legislation can leverage Quture’s extraordinary advantage when CMS transitions from billing to clinical data.
One observation at the conclusion of our discussion of QualOptima operating on 2nd generation data lakes is important to understand why peer review has not achieved the goals of quality and risk management. Physicians resist the punitive use of performance and outcomes data, which is why peer review has not provided solutions sufficiently to improve quality and prevent medical errors. However, when insight from analysis using health data science is available to physicians, retrospectively and in real-time to improve clinical processes and outcomes and to achieve value-incentivized care, quality and risk management will be welcomed as clinical decision support.
Quture engineered our QualOptima Health Informatics Platform to empower healthcare organizations to eliminate functional and information silos that exist in traditional quality management, risk management, and credentialing by the medical staff office. The QualOptima common data platform shared across those traditional boundaries as Value Data Centers is the future of using 2nd generation data lakes retrospectively and in real-time at point of care. When Quture transitioned our technologies to Health Data Science, healthcare organizations become empowered to provide data-driven knowledge solutions. These solutions are based on efficient and effective assessment of the impact, cost, benefit and risk of continuously monitored optimal clinical processes in the framework of human and contributing organizational factors.