Real-time intervention

Real-time intervention in health care has been attempted before with poor results. What makes us think that we can do better this time?

Well, to start with, we have learned from previous experience that inundating clinicians with false-positive information and information that they already know but does not change their clinical practice is a no-no (see drug-drug interactions).

Clarifications of their documentation to support better billing and quality reporting, however, are seen as non-intrusive to their clinical care but improves the administrative practice.

Although clinical decision support is a requirement under Stage 2 Meaningful Use.
According to HealthIT.gov, and includes tools such as “computerized alerts and reminders to care providers and patients; clinical guidelines; condition-specific order sets; focused patient data reports and summaries; documentation templates; diagnostic support, and contextually relevant reference information,” these systems are not likely to be used unless the False Positive Rate is significantly reduced based on real clinical experience.

Only through machine learning, prescriptive and predictive analytics can the sensitivity of the systems be improved. Translational research has been used to apply new learnings from retrospective database research. The CDS Consortium, at the Vanderbilt has identified “knowledge translation and specification” as an important research objectives. In addition, the Consortium for Healthcare Informatics Research, a VA-based research program located at the University of Utah is also advancing many of the methods required to conduct better predictive and prescriptive analytics.

In the meantime, real-time alerts can help providers avoid mistakes and improve quality by focusing on a few set of prioritized reminders and notifications related to Performance Indicators which have been studied extensively by AHRQ, JCAHO and CMS, amongst others. Core Measures and Patient Safety Indicators are 2 general domains which are important to notify providers of relevant information.

Blueprint Accelerator Boosts Hiteks

Hiteks entered the Blueprint Accelerator Program in July 2014, from which it received a small financial investment but a large emotional investment. What I mean is that the Blueprint Fund and its Management, a group of young Ivy-league health care entrepreneurs, shook up Hiteks’ approach to sales and revenue-generation. Their goal was to show the founding members of a start-up organization how to focus on a repeatable sales methodology which would validate (or not) the current business model.

Up until Blueprint, Hiteks had been engaged by health care organizations interested in better accessing either their own data or real-world data that Hiteks had the right to de-identify and share. Clinical NLP technology was in its early phases, most people thinking that its true value was in retrospective analysis of Big Data to produce population-level insight into quality, better-disease understanding, or to analyze risk. Although these have certainly proven to be important areas of application of the technology, they all have a limited (and competitive) nature to them since they don’t impact clinical care or workflow.

What we found by talking endlessly to all of the Blueprint alumni and network is that there was a muted interest level when it came to retrospective analysis of data. Not that it wouldn’t be helpful, but a software company needs to produce code to conduct repeatable processes, not one-off questions and poorly defined hypothesis-driven search for answers. Most of the NLP companies that Hiteks compared its technology to at the time fell into the trap of designing their systems to facilitate such retrospective analytics. The Blueprint management urged against any technology marketing and sales that couldn’t be repeated 1000 times by the same application and where demand was greatest.

That’s when Hiteks realized it had to apply its technological know-how to the real-time decision-making of the clinician at the point of care. This was the only step in the health care delivery process which allowed for the validation of data points to become part of the medical record. It would have to be fast enough for clinicians to quickly make a decision on the content of the NLP output so that their workflow would not be hindered. It would have to be integrated into their existing documentation system, the EHR, so that they wouldn’t have to use yet another system to give them advice.

Once the real focus of Hiteks was realized and agreed upon by its founders, communicating the value of the technology and the company was the other area that was developed and shaped with the help of Blueprint. Once we got the email lists of prospective leads organized from multiple sources, we sent out blast emails which got the 3% traction/response that we were told to expect. This meant that we found the early adopters of our solutions, and we could focus on selling to them the value proposition we created. This helped us get our first 3 health system clients for a total of 50 hospitals and affiliated clinics signed up to use our software. The rest is history…

Archives

Contact us for information on our products and services.
For those who schedule a demo, we offer a complementary yoga mat to promote good health and wellness.