Divya Malhotra, director of analytics and innovation at Connecticut-based Stamford Health, spoke with Becker's Hospital Review about deploying data to clinicians and seizing the power of descriptive analytics.
Question: How has the data analytics program at Stamford Health developed over time?
Divya Malhotra: We're running about 100 data sources, which includes clinical data, financial data, patient satisfaction data and our state association data, so we have analysts deploying reports across the organization. Over the past three to four years, we have connected all of our data sources to visualization software like Tableau, which has brought out the power of the data that we already had before. We can now produce analytics across the system, which affects our patient care and quality management, because our processes are empowered with better data and better analytics.
Q: What is the biggest challenge to maintaining a data analytics program?
DM: One of the key questions is, "how do you achieve that tipping point?" Everybody wants good reports, everybody wants good data, but the challenge is that not everybody is quantitative by nature, not everyone wants to look at sheets and sheets of raw data. The challenge, really, is that you can have all the back-end in place, with all the raw data, but how do you deploy it to your clinicians? How do you present the data so that clinicians can pull their own reports every day, without having to deal with processing it themselves?
There is that optimal zone; you don't want to give them too much self-service, because they're busy doing their own day-to-day work. You need to provide them with enough analytics power to see data in a couple of different formats, but to steer clear of giving them an overload of information.
Q: What types of patient care and quality initiatives have grown out of the analytics program?
DM: We started by collecting all of the manual data from the EMRs, and are now creating an enterprisewide quality scorecard, where we are connecting Excel spreadsheets and hooking it up to our data software. That gives us one centralized place for all of our quality measures, which helps us to analyze and monitor quality across the organization.
As an example, this data showed us that we had a higher turnaround for our lab tests for patients in the ED than we wanted. We realized that one part of the reason was because a lot of samples were getting hemolyzed, so clinicians had to go back and redraw the blood. We looked at the data, did some root cause analysis, made some improvements to reduce the hemolyzed samples — like using straight stick needles, instead of butterfly sticks, to draw blood, which reduced the hemolysis. And because these straight sticks are cheaper, we actually had savings as well as increased efficiency. We continued to monitor our lab turnaround in the ED, and saw that it actually went down by 19 minutes.
Q: Moving into 2017, what do you see as the biggest area for growth in data analytics?
DM: One of the areas that has come up a lot is patient data in mobile devices. In the next few years, I think you're going to see more data being collected in mobile devices, so that's where we should be looking. That way, you get real-time data from patients on their health status, and you can really drive that to improve outcomes and to improve quality.
But, with that said, I want to make the caveat that descriptive analytics are not going anywhere. Even though it's very simple, and it's not the 'next big thing' like prescriptive analytics or predictive analytics, you're always going to want to look at and learn from your historical descriptive data. We always have to look back and learn from the past.