Improving Healthcare (and Pharma’s) Performance through Analytics
Leveraging data through analytics can improve healthcare services and health outcomes for patients and communities.
Pharma has been dealing with major talks about Big Data in recent years. Some healthcare organizations, however, don’t consider themselves handlers of Big Data yet. According to Seattle Children’s Hospital’s (SCH) Chief Data Officer, Eugene Kolker, “In our case, we’re not Amazon, Microsoft or Wal-Mart. We are not in Big Data, really. We call it ‘many data’ or ‘complex data.’ We have a lot of different data types and they’re interrelated and connected in a lot of different and complicated ways.”
Organizations such as SCH may not be in Big Data yet, but they still handle very diverse sets of information, ranging from scientific findings to sensitive patient details. Sorting through the clutter of complex data can, nonetheless, be an enormous task to undertake. Through proper analytics, data can be used as raw materials in coming up with a well-founded decision. “Basically, we leverage data through analytics as our strategic assets to primarily focus on improving care and safety of our customers who are patients and families,” he says. The approach developed at SCH is called ‘PPT’, meaning: People – first, Process – second, and Technology – third, all empowered by Data and Analytics. (For more explanation and examples of this approach please visit cdoanalytics.org). “This is a consistent and effective approach which helps our clinical and business leaders to make better-informed, data-driven decisions and follow them up with prioritized execution plans”.
Improving management and effectiveness of an organization
One of Kolker’s main activities is called the Benchmarking Improvement Program. The program is aimed exactly at improving effectiveness in the customers’ care and safety. “It is basically an in-depth benchmarking of our performance now versus our performance in earlier years, and also versus top performers industry-wide,” he informs. “Physicians, administrators, managers, and nurses within business units and across the entire network want to learn what the best practices are industry-wide. They definitely want to compete with the best performers in the nation,” says Kolker. He adds that health professionals are eager to learn how to best handle uniquely challenging medical situations such as infections, for example.
Providing health professionals with the catalyst to improve their performance is the single goal of improving patient outcomes. “The ultimate focus of the whole [benchmarking] program is to improve what we do, to be more effective in how we do it, and to deliver superior service to our patients and families,” he says. “And the PPT approach enables this by empowering clinicians, managers, and staff.”
Having a single goal unifies the different health professionals within the organization and encourages collaborative work. “We focus primarily on clinical outcome measures. We are basically involved with a lot of people from multiple service lines. That’s the whole point of this program. It’s not just one or another service line or business unit. It goes across the board.”
Improving clinical outcomes
Analytics doesn’t only describe and measure health outcomes, but can also result in prescriptive information. Kolker cites the SCH’s Pulmonology’s flu vaccination service as an example. “They’ve made a lot of improvements based on analysis. We worked together on which of those analyses can be actionable and how. They’ve improved their service and performance tremendously compared to a year ago, let alone two years ago. Now the majority of their patients are getting vaccinations,” he shares. In turn, these will translate into improving the health status of patients, reducing admissions, and reducing complications.
Facilitating medically challenging problems
Improving healthcare delivery in certain departments can be more challenging than others. Kolker cites the management of infection in the Intensive Care Units as an example. “ICU infection rates are a major area for improvement. We did improve a lot last year, but this is a never-ending battle,” he says.
Kolker also observes similar situations in other healthcare organizations. “I was recently at a conference and one of the keynote speakers before me was the Chief Nursing Officer of another hospital system. She pointed out, ‘Every morning, I start with the same task I started yesterday’”.
Managing infections in a children’s hospital is critical. “Kids are under a lot of pressure health-wise,” explains Kolker. “Reducing infection rates is a major factor in their well-being. We need to continue to learn the best practices because there are a lot of improvements and innovations across the country and the world.”
As a result of their benchmarking program, the hospital can determine which organizations to look to for best practices. According to Kolker, “We have inspiring meetings within walls. Then our medical and administrative leadersare reaching out to those with the best performance. Our benchmarking program allows them to know who is doing better than they areand in which specific areas.” He notes that health professionals are really collaborative in learning and sharing best practices across organizations when it comes to patients.
Choosing analytics tools and defining engagements
There are different and appropriate tools available to analyze data, but choosing which one to adopt depends on the project’s needs. Data and analytics work requires usingtools in conjunction with each other. “We utilize regression, forecasting, and optimization models,” he says. “We really are big fans of R, which is an open source platform for statistical analysis. In many cases R is the best for analysis and modeling optimization. We also utilize other systems including internal EDW (Enterprise Data Warehouse), Jawa, Python, Hadoop, or Tableau and D3 for visualization needs. But Technology comes third after People and Process!”
The choice of tools is largely based on the objectives of the project, and successfully setting the objective is dependent on whether the specialists involved in the project are willing to collaborate openly: People – first! Starting from your ultimate customers and their needsand then following with business challenges and opportunities is key. Before any project commences, both the participation and expectations from other parties should be clearly defined. “We have very clear processes of engagement,” explains Kolker, referring to the ‘process of engagement’ flow defined on their website. It visualizes the timeline of the project and at which points thedata and analytics team would need to get in touch with parties in the project before proceeding to the next step: Process- second!
More importantly, it paints a picture of how important everyone’s participation is in ensuring that all pertinent data and expertise are made available. “It’s clear what our clients bring to the table and what our team brings to the table. We’re trying to get a 360-degree view of the problem we’re analyzing. Say we figured out what problem we’d like to address and we know more or less what kind of data we get within walls. At that point, we look outside the walls as well. We feel that it’s great and it’s possible to bring other data sets to complement what we have within walls,” says Kolker.
Working closely withparties who are eventually going to make useof the analysis is critical in what Kolker calls the ‘translation’ step. “Most people in healthcare didn’t come to healthcare to work on analytics. They came to healthcare to help other people. So, we need to translate what we’re able to develop into something very clear, very succinct, and actionable,” he adds. “In turn, they will become FDAs –Friends of data and analytics”.
Analytics and innovation in pharma
Leveraging data through analytics also helps improving data accessibility for other professionals and facilitates innovation in healthcare and biomedical research. Kolker also designed a research resource called the Multi-Omics Profiling Expression Database (MOPED), which allows researchers to browse gene, protein, and pathway expression information in humans and model organisms from publicly available studies.
MOPED benefits many scientists, researchers, medical professionals, and physicians by enabling them to rapidly browse and compare seemingly unrelated results to supplement their own work. According to Kolker, “Our favorite thing is when we get emails from people thanking us for connecting the dots for them in experiments done in different labs and different continents on something they plan to work on or have already worked on-emails like, ‘Now I understand why our results were like this,” or ‘Now we know what we are going to do next.’ This is really empowering other researchers, including those in pharma, as we can see from the MOPED’s usage.”
Kolker believes pharma gets plenty of unique and useful data available from public domains, academia, and non-profit organizations, which can be complemented with adequate analytics expertise. “Leveraging these with other, unorthodox data types will enhance pharma’s unique strengths. Imagine what next improvements and innovations are going to come from that leveraging, especially, if this will be done in a smart, effective, and very purposeful fashion. Clearly, pharma can put strategic thinking and serious resources towards doing that,” he adds.
Pharma can develop a competitive edge through successful data and analytics that result in data-driven decision-making and execution. Kolker’s suggestion? “Consistent and effective approach to leveraging many complex data through analytics can become one of pharma’s competitive advantages”.
Since you're here...
... and value our content, you should sign-up to our newsletter. Sign up here