Experts: Big data need big direction
Patient-reported data can enhance comparative effectiveness research and provide better outcomes for patients if the data are aggregated and available to researchers, according to investigators at Duke University and Duke Clinical Research Institute.
In a paper published in Health Affairs, Amy Abernethy, MD, PhD, professor of medicine at Duke University, and colleagues explore patient-generated data and how this unique information can be made useful in both clinical and research settings.
In an interview with HemOnc Today, Abernathy said an enormous amount of data is being collected by patients, particularly through the use of EHRs, but the data needs to be accessible to researchers and patients to improve the quality of care.
“There is this big conversation about big data, but big data is sometimes disconnected from patients, and I want to make it more connected to the people we serve,” she said in the interview. “No matter what kind of data sets we’re talking about, one of the frequently missing components is the patient’s story and experience — their symptoms, quality of life, satisfaction of care, and whether they thought treatment was worth it. That information has to be sourced from people themselves. We have to be intentional about getting that kind of data set directly.”
What’s more, data on patient experience is important in evaluating the role of services.
“If we’re going to make patient care more patient-centric, by definition it needs to incorporate the kind of outcomes that matter most to real people, such as having the quality of your life get better. In order to know if the quality of life got better, we need to understand what quality of life looks like, and not have a bunch of proxies,” she said.
As an example, Abernethy said that if a cancer center notices that a large number of patients are self-reporting high levels of pain, then that center may want to establish a pain clinic or treatment protocol to address the issue.
However, she warns that the right questions need to be asked of patients. She said that if every patient was asked if they were in pain and to rate the level, but the data were out of context of what else is going on (like surgery or a background history of migraine headaches), the data might become too convoluted.
She is, however, optimistic about the future integration of patient-reported data into the larger system of health care reporting. She said she is involved with the Patient-Centered Outcomes Research Institute (PCORI), which has funded PCORnet, a patient-centered network of 18 patient-driven research networks and 11 clinical data networks. The goal of the program is to develop an infrastructure of research data on a large scale, and to pull data from EHRs and other resources. The NIH Collaboratory was developed to integrate electronic health and clinical trial data, and she said other resources like the ASCO CancerLinQ will provide access to cancer data.
Moving forward with data integration and accessibility will be challenging, however, because data is buried in many disparate places, from physician’s records to data on smartphones.
“The first task is to say that this is important,” Abernethy said. “The next part is tactical, but so many things are starting to push this idea forward that I am confident things will start to align. We’re starting to have the social conversation about the importance of collecting information directly from patients. In response, we are seeing a shift in focus towards patient-generated data that is more in line with public sentiment. To give this momentum we’ve got to be able to put the vision out that people can start to understand.” – by Shirley Pulawski
Disclosure: See the full study for a complete list of relevant financial disclosures.