Turning Predictive Modeling into Practical Medicine

Oct 03, 2014 at 06:33 pm by Staff

“Predictive modeling, on its face, has absolutely no value.”

That’s a strong statement from a man who makes his living in the field, but Dan Hogan, president and CEO of Nashville-based Medalogix, is firm in his assertion.

Not surprisingly, though, there is more to his sentiment. “Telling risk does nothing if you don’t facilitate a change,” Hogan continued. “Where we started as analysts, we’ve become analysts and clinical behavioral consultants.” It’s that ‘and,’ he said, that makes all the difference to patients and the providers caring for them.

Hogan, a former home health agency owner, founded Medalogix in 2012. “We provide predictive modeling and workflow tools to home health providers, and in a secondary market, to hospice providers,” he said. Hogan added he believes his company to be the only predictive modeling toolset specifically designed for the post-acute care market.

While the analytics are certainly not limited to senior care, Hogan said the sphere in which Medalogix operates means their work typically impacts older patients. For example, he noted, the average home health patient on his census is 82 years old with two chronic conditions. “Approximately 60-65 percent of the patient records we’re analyzing on a day-to-day basis are Medicare beneficiaries … 30-35 percent are private pay,” Hogan said. However, he continued, “Well over 95 percent of patients we analyze data for are Medicare-qualifying in age.”

Using complex algorithms, Hogan’s team is able to assess which patients are more likely to be readmitted to the hospital within 30 days or to predict patient mortality within 180 days. “On the back end of that analysis, we provide the workflow solutions that facilitate the reduction of the risk of either of these terrible outcomes,” he said.

Risk, he continued, is not singular. Not only is it dependent on how sick a patient is but also on the skill sets of the clinical organization caring for that patient. Congestive heart failure is certainly a risk factor for both hospital readmission and mortality, but Hogan noted a patient’s risk of those outcomes is inherently less if they are receiving care at a center of excellence. “Not everybody is good at everything. Clinical capability really does matter,” he said.

In determining risk, the company’s data scientists customize the algorithms taking into account a client’s specific concerns, patient demographics, organizational information and historical data. Again, Hogan stressed, the resulting data insights are only as valuable as the actions they inform.

And those actions can and do change the trajectory of risk. Drawing from his own background as a home health agency owner, Hogan recalled, “We were terribly ineffective at managing bed sores.” Trying to reduce the patient risk, he sent three nurses to a training seminar. “Within 90 days, our rate of readmission for decubitus ulcers was down to 4 percent.”

Similarly, Medalogix uses data to determine optimal transference to hospice care. “At a very high probability of mathematical accuracy, we identify patients at elevated risk for mortality,” he said. That knowledge, he continued, is very powerful in the hands of experienced clinicians … and powerful for family members, as well, when it comes to making informed decisions.

“Identifying a patient that is hospice appropriate earlier in that end-of-life cycle tends to result in longer life,” he pointed out. More importantly, Hogan continued, those patients also enjoy a much higher quality of living during the time they have left.

While the data analysis is highly accurate, Hogan is the first to say there are always exceptions to the rule. “Intuition trumps data,” he added with a laugh … at least when that intuition comes from experienced clinicians. “There are always people we have ranked very high on risk, and the clinician will say, ‘Nope, you’re wrong on this one.’ They are usually right.”

An indomitable spirit can certainly carry a patient past mortality expectations when there is a deep desire to see a grandchild married or a great-grandbaby born. For most patients, though, predicting readmissions or mortality is extremely difficult when relying on intuition alone.

“The analysis of clinical data is providing an opportunity to glimpse into the future,” Hogan said. “In healthcare, historically the quality and clinical decision-making largely depended on the experience and skills of good doctors and nurses. Now, Medalogix and other companies are building analytical tools providing data to support those experienced doctors and nurses, making data-driven decisions possible.

“We would absolutely say medicine is both a science and an art,” he continued. “Having both the data and the artistic insight maximizes resources and is better than just one or the other.”

Clearly, that message is catching on with clients as the landscape shifts to pay-for-performance models. In the last year, Medalogix has increased its client base by more than three-fold and doubled staff to keep up with demand. The company recently received a $5 million cash infusion from Connecticut-based Coliseum Capital Management to help fuel this continued growth.

Sections: Archives