Big Data from the Healthcare Provider Perspective
eyeforpharma speaks with Prasanna Desikan,Ph.D, Senior Research Scientist at the Division of Applied Research, Allina Health, about how providers are increasingly using insights from big data analysis in their treatment of patients.
‘Big data,’ the term for the collection and analysis of the increasingly large amounts of information produced by today’s ever-connected world, has been talked about as an emerging trend that will change the way pharma does business. Yet how does the healthcare provider view this phenomenon?
Prasanna explains that “big data in healthcare is a new phenomenon. Big data analytics have existed for a while in other domains, but in healthcare it has been a relatively new development. Some work has been done by payers with access to patient data, but from a provider perspective, the gathering of data on Electronic Medical Record (EMR) systems and the analysis of that data has only recently started to happen. One of the key reasons big data analysis in health systems is relatively new is that data collection, storage and access technologies have only been recently implemented in most healthcare provider organizations. Before this there has never been a consistent effort to develop ways of collecting, storing and accessing data, meaning that there hasn’t been a foundation for any kind of big data analytics; to implement these kinds of analytics you need to have strong patient datasets as a foundation, as well as the ability to store and access these datasets easily.”
The ability to analyze large amounts of patient data is of great benefit to healthcare providers, according to Prasanna. “What EMR systems have done is to allow us to collect data collaboratively and to integrate it into one uniform format so we can see a bigger and more integrated picture of that data. This gives us greater knowledge than if we were looking just at individual patient or clinic data. If there is a patient with a particular condition being treated by a certain provider, there may be patients with the same condition who have been successfully treated in the past. On the other hand, there may have been other treatments tried with this kind of patient that haven’t worked so well. With big data you can start to see trends, what works well and what doesn’t work well in certain situations - understanding these kinds of patterns can be really helpful.”
Attention needs to be paid also to the way the data is stored and accessed, in order to enable successful analysis.
Some research organizations, particularly within the field of oncology, are analyzing big genomic datasets to isolate patient subpopulations likely to respond better to certain interventions. Yet Prasanna sees these datasets as covering “only one aspect of care delivery.” EMR data, on the other hand, “gives you information on blood pressure, pain, and other things that are not available from genomic data. We cannot treat all patients purely based on genomic data. Clinical and socio-economic information, demographic information, financial information, all of these play a very important role in the provision of care - it is a different dimension of knowledge.”
However there are some challenges with collecting and analyzing data in a clinical setting: “working with big data needs quite a lot of resources. To benefit from big data you need to be collecting the right data, to have the right EMR system, and to manage the workflows of the clinicians collecting data. Attention needs to be paid also to the way the data is stored and accessed, in order to enable successful analysis. Once all this is done, what do you do with the analysis? You take it to the different teams, such as financial and operational teams in your organization, who relay the information back to the clinics. This eventually leads to feedback on the kind of care that the organization provides, as the goal of carrying out this analysis is to improve the overall health of the population.”
There are a lot of challenges involved in figuring out what kind of data to enter, how to collect it, and when to collect it.
Prasanna sees one of the biggest factors in determining the success of a big data project is how it is incorporated into the workflow of the practicing clinicians. “Many organizations are just now implementing EMR systems, and are learning how to incorporate them within their workloads. There are a lot of challenges involved in figuring out what kind of data to enter, how to collect it, and when to collect it. These kinds of issues are still being worked out, and this uncertainty will continue for a few more years while EMR systems and the processes around them mature.”
Prasanna gives an example of how the analysis of big data has been put to practical use in Allina hospitals, in order to avoid potentially preventable events (such as hospital readmissions) occurring in its patient population. “We identify the risk for patient based on patterns we observe from the historical data. By looking at the patterns of people having potentially preventable event, we can work out the rules to identify future patients at risk. When a person is identified as at risk of a hospital admission or other event, a care team will engage with the patient and encourage them to enroll in a care program. The patient can of course decline this, and we have protocols on how to approach the patient. The patient must understand why we are approaching them, and they must be given the ability to choose whether they want this care.. This way, you can reduce the burden on the system caused by these events.”
We are not yet at the stage where we have started including big data analysis in decisions over which drugs make it onto the formulary.
Prasanna describes how Allina has already seen “a difference and improvement in terms of reducing hospital readmissions using this method.” He thinks that health care providers have “a unique opportunity to go through the whole data-gathering process from end to end. At Allina we are able to apply the lessons we learn from big data to the care of the patients.” Prasanna judges the success of a company’s big data projects not only in clinical endpoints achieved, but also in how widely the project is accepted by those who use it in the front lines of care. “We measure metrics such as whether the number of readmissions have been reduced, but in addition to this, we also measure whether the tool is actually being used. A great way of looking at tools used to improve care is to ask how well they are accepted within your organization by the people who actually deliver the care.”
While big data is being used in Allina to better manage their patients’ healthcare, and to save costs, he acknowledges there are some areas where the insights of big data are still not being put to use. “We are not yet at the stage where we have started including big data analysis in decisions over which drugs make it onto the formulary. We have thought about it, and we will probably be doing it in the future, but like most local healthcare providers we are not quite there yet. I would speculate that the reason for this is that healthcare providers are still grappling with their EMRs, understanding how to implement them, and getting the basics down in terms of the delivery of clinical care. Organizations probably view incorporating the analysis of big data into formulary decisions as a next step, but first we have to get better at using basic EMR data to manage clinical care.”
Prasanna Desikan will be joining other big data innovators later this year at the annual Real World Evidence conference in Washington DC. For more information on Prasanna's presentation or to find out who else will be attending visit the official website.
Since you're here...
... and value our content, you should sign-up to our newsletter. Sign up here