After hearing so much about the promise of big data, it was refreshing to hear about how this “new” strategy works in reality.
During a recent O’Reilly webinar about personalized medicine, Arijit Sengupta of BeyondCore explained how one of the company’s clients — a large healthcare system — analyzed patient length-of-stay patterns.
The project involved:
- 247,000 patient outcomes
- 534,000 variable combinations
- 9 million calculations
With the Rise of AI, What IP Disputes in Healthcare Are Likely to Emerge?
Munck Wilson Mandala Partner Greg Howison shared his perspective on some of the legal ramifications around AI, IP, connected devices and the data they generate, in response to emailed questions.
Sengupta said there were 35 key insights auto-detected by Lucid, the company’s business intelligence software. He highlighted three:
- For the exact same surgery, patients in one hospital were recovering 14.5 days earlier than patients in another.
- Patients with “altered mental status” had 7.5 days excess length of stay for surgeries across all hospitals.
- Patients with one particular insurance who received a certain treatment were staying 20 extra days.
Sengupta said this analysis made the extent of the problems much more clear.
“Most of the time, executives don’t have the information at their fingertips to identify these problems,” Sengupta said. “They can’t collaborate within the organization to find the solutions.”
A Personalized Approach to Medication Nonadherence
At the Abarca Forward conference earlier this year, George Van Antwerp, managing director at Deloitte, discussed how social determinants of health and a personalized member experience can improve medication adherence and health outcomes.
Sengupta said BeyondCore takes a search-oriented approach to statistics to provide “advanced analytics for all.” This lets users ask questions without writing queries.
“We had to come up with our own proprietary approach, but you could replicate this in any system and focus on examples instead of statistics, for example,” he said. “The principal is more important than the code.”
Sengupta said the executive dashboard uses tiles to represent reports.
“You flip the tile and it shows more detail for each example,” he said.
One audience member asked how an analyst could be sure the computer is doing its job given the visual approach to reporting.
To make sure his team could win debates with statisticians, Sengupta said the system also allows a business analyst to see the variables and see each examples for each variable combination.
“They can see the regression and determine what is R-squared as well,” Sengupta said.
Today is the second day of O’Reilly’s Strata conference that includes some of the other speakers in the webinar. You can watch the event and check out the agenda here.
