Predictive Analytics and the Human Side of the Equation

Predictive analytics…we have all heard the term, and it’s one that tends to grab our attention today in the healthcare industry.  And for good reason: who doesn’t want a tool to help us predict clinical and business outcomes that would help improve care and satisfaction in this ever-evolving and accelerated industry?  It’s exciting to consider predictive analytics as a way to link financial reimbursement to clinical outcomes.  And in our reforming healthcare industry, this technology and intelligence must be in place. 

But let’s also not forget the human side of the equation in predictive analytics.  Even though technology provides us with data-driven tools to help us predict our business and clinical outcomes, we must still maintain human judgment and recognize that a tool should cannot replace our own clinical judgment.

I recently read an article published by Healthcare IT News on personalized medicine, noting

  • 2/3 of healthcare organizations believe personalized medicine is already having a measurable effect on patient outcomes, and
  • 65% of respondents called for an increase in predictive analytics. 

However, we need the right tools in place to support the governance and technology required to make this a reality. 

Our health IT community can help by starting to invest and support this need for useful and usable tools and technology.  I feel this technology will become more sophisticated and tailored to the individual; but not all models will ever be ‘one size fits all.’  Clinical judgment will always be important in the care of our patients.

A few years ago, a nationally recognized transplant center implemented a patient-based risk tool for use in its weekly ‘selection meetings’ to help clinical teams look at the patient’s risk factors and predicted patient graft and survival outcomes.   Up until this point, the transplant candidate’s current medical condition is presented, and a group of multi-disciplinary clinicians, who engage in conversation discussing risk factors, make a decision on whether or not to place the transplant candidate on a waitlist.

During the selection meeting and introduction of the risk-assessment tool, the clinical team became somewhat challenged because they were immediately reviewing the risk-assessment tool to select the transplant candidate, versus relying on their own clinical knowledge.  Keep in mind, the predictive analytics tool and relevant data conveyed the message that technology overrode their own clinical expertise in analyzing how these patients would clinically rate.

Here, we have highly educated, trained and clinically credentialed providers suddenly so focused on visual the predictive capabilities of the technology’s graph and charts instead of applying their own human judgment on other factors for that individual patient, factors unique to that patient and the specialty field. 

After learning how predictive analytics could assist instead of override their judgement, the clinical team modified its decision-making process and used the tool as ‘validation based on risk’ f to waitlist, or not, a patient for a transplant. 

Even in cases where the predictive analytics tool did not support their decision, the clinicians experienced some ‘aha’ moments. The tool still gave value to the case by identifying factors, such as those related to patient compliance and lack of family support, both important for a successful clinical outcome in a transplant patient. 

Humans are complex, and patients are individual.  We must not lose sight of that.  But we also know, in health IT, we can and must continue to work to make predictive analytics a very useful and integrated tool for decision making to support successful outcomes.

Learn more about blending human factors  with technology to optimize outcomes from a HIMSS 3-part series, “The Integration of Clinical and Business Intelligence into Clinical Workflows” (Part 1 – Introduction; Part 2 – Case Studies; Part 3 – Future Scenarios).

Have you ever had an experience where technology and clinical judgment play a role?  Have you thought about how we can better anticipate and prepare for how we utilize predictive analytics for better outcomes? 

Join the HIMSS C&BI Community today to learn more about predictive analytics and other critical topics that will help you on your journey to Turn Data to Action

Predictive Analytics; Workflow; Human Factor; Clinical & Business Intelligence