Guide to business intelligence and health IT analytics
A comprehensive collection of articles, videos and more, hand-picked by our editors
Most healthcare organizations possess some degree of business intelligence or analytics capability, but until clinicians can use embedded analytics at the point of care, the full potential of real-time information at the fingertips of doctors and patients will be difficult to realize.
It would be challenging, not to mention highly inefficient, to operate a modern healthcare organization without even basic business intelligence (BI) functionality for decision support. In typical organizations, this capability likely consists of data from source systems being stored in an enterprise data warehouse or other storage repository. Analytics, BI and reporting are provided by one or more special-purpose tools.
Healthcare business intelligence tactics
Radiology can't afford to ignore BI
BI can create more custom reporting
Data analytics and BI quiz
This configuration of BI and analytics may be suitable for most business management decision-making requirements, as well as research; however, this arrangement offers little, if any, benefit by way of analytics to healthcare providers working on the front line. Clinicians must assimilate large quantities of information, including test results, best evidence, and their own clinical judgment during the course of treating patients. Unfortunately, the information systems that are currently available to most clinicians do not assist them very well with this task, and poorly designed systems may only obfuscate the process.
Initial efforts to aid clinical decision making consisted of tools that would enable clinicians to find the best evidence online during the course of treatment. The reason more analytics are not available at the point of care is because it can be challenging to integrate analytics into a clinical environment, such as an electronic medical record (EMR). Therefore, clinicians may require workarounds such as a push to mobile devices or some tool other than the primary clinical system in use.
Although this hybrid approach may be sufficient for some clinical uses such as alerts, this strategy likely cannot support most real-time clinical decision-making needs. The technology is advancing, however, to the point where it is becoming feasible to embed more real-time analytics in the clinical systems, such as EMRs, that clinicians are already using at the point of care.
Embedded analytics integrates workflow at point of care
There are many potential benefits for healthcare providers in bringing the power of analytics to the point of care.
- Synthesis: Embedded analytics will enable the rapid integration of multiple sources and large volumes of data. For example, such tools will quickly determine the relevant and best outcomes and different courses of treatment for patients with similar conditions. Embedded analytics can also inform clinicians in real time on how their performance stands given the decisions, clinical outcomes, and quality performance of their peers.
- Workflow: Analytics embedded within clinical applications are better integrated into existing workflows so there is no need to look at other information sources that disrupt the process of clinical decision making and caring for patients.
- Context: When embedded inside clinical applications, analytics-generated insight can be presented alongside and in context of patient data.
Challenges of embedded analytics
Although there are definite tangible benefits of embedded analytics within clinical applications, there are technical and other challenges that are preventing the widespread use of embedded analytics in health IT. A few of these challenges include:
- Integration: Making a seamless analytics transition is a major technical challenge. The usability of analytics at the point of care will be impacted greatly by how well analytics is integrated with the user interfaces clinicians are used to seeing. Point-of-care analytics needs to be patient-context sensitive, and must provide the necessary insight clearly yet unobtrusively to the patient visit.
- Relevancy: Embedded analytics must anticipate the questions clinicians will need to have answered. In doing so, the information must be targeted and specific to the needs of the patient and care delivery. This cannot be random insight generated by a machine-learning algorithm.
- Clinical competency: A very real barrier is the perception that embedded analytics will diminish providers' clinical autonomy and have them engaged in "cookbook medicine" -- or worse, that their clinical judgment will be second-guessed by computers and perhaps even overruled.
Expanding analytics in electronic records
Clinical system vendors are making great strides to ensure electronic health record systems are user-friendly and intuitive. The next steps are for vendors to expand the use of analytics within the clinical systems, and not just as an add-on decision-support (or reporting) package. In future articles I will be profiling examples of where, in my opinion, embedded analytics is done right.
About the author:
Trevor Strome, M.S., PMP, leads the development of informatics and analytics tools that enable evidence-informed decision making by clinicians and healthcare leaders. His experience spans public, private and startup-phase organizations. A popular speaker, author and blogger, Strome is the founder of HealthcareAnalytics.info, and his book, Healthcare Analytics for Quality and Performance Improvement, was recently published by John Wiley & Sons Inc.