It has long been a goal in medicine to extract not just data but real intelligence from patient files. In the oncology field, we are finally seeing that goal become reality.
The automated clinical annotation of electronic medical records, together with careful curation of genetic test results, is transforming the type of information available to healthcare providers. For example, big data advances have made it possible for oncologists to visualize a patient’s complete healthcare journey without having to create a mental map based on the many disparate sections of an EMR.
Extracting intelligence from each patient’s records is also enabling a better understanding of health at the population level. Oncologists can use sophisticated analytics to see at a glance how any individual patient’s clinical profile maps to a tailored cohort. That same intelligence is making it possible to more precisely match patients to clinical trials, as well as to assess population-level priorities, including quality of care and research.
Today, too many physicians struggle to work with EMR systems that were never designed to provide clinical intelligence that could improve patient care and outcomes. As newer, easier-to-use, and more effective data tools become available, hospital systems and medical practices have the opportunity to support oncologists and help them make the most of patient data.
State of the Data
It is no secret that the process of accessing patient records, entering data, and trying to extract knowledge from EMRs is fraught and painful. The platforms used today for patient data were originally designed for billing purposes, not for clinical care. No wonder, then, that they fail to offer an intuitive user interface for physicians who simply want to understand a patient’s healthcare journey.
Preparing for a clinical interaction with a patient requires oncologists to log into the EMR and then click through many sections, each storing disparate pieces of information. Data that would ideally be readily accessible in a single dashboard — patient demographics, stage and type of cancer, histology data — is instead hidden away in so many silos. Valuable notes from previous visits with other physicians in the same healthcare system are similarly difficult to find, and are not optimized for searching. The platform is designed for physicians to enter and review patients’ data without standardized vocabulary, leading to the burial of information in a mountain of unstructured physician notes. Test or imaging results may be locked away in PDF format, requiring the user to open each file in order to see what’s there. And important data from a patient’s meetings with doctors at other practices or hospital systems will be completely inaccessible unless that patient has taken the time to request the information, which is often shared via obsolete storage devices such as CDs.
Change Is Coming
Twin trends, driven by technological advances, are ratcheting up the pressure on health data systems and simultaneously offering a solution.
The first trend is the massive increase in the amount of data being generated for each patient. Thanks to innovations in DNA and RNA sequencing as well as liquid biopsies, the molecular profiling data available for anyone with cancer is growing exponentially. Where we once had single-gene readouts to tell us whether a patient’s breast cancer might respond to a targeted treatment, we now have longitudinal data covering the entire genome of both the patient and her tumor, how the tumor DNA or RNA changed in response to treatment, and real-time information about metastasis. In addition, patients are bringing their own data sets to the table: wearable sensors, phone apps, and other consumer-oriented tools are all generating useful information about a person’s activity, environment, diet, sleep patterns, and more. Never before have oncologists had access to such comprehensive views of a patient’s situation. But conventional EMRs cannot keep up with this explosion of data and the enormous files they entail.
Fortunately, the other trend — based on progress in handling big data — is sweeping in with a fix. Natural language processing, for instance, has made sufficient progress to support semi-automated data extraction from EMRs for more clinically focused intelligence. This can be used to structure and harmonize data, making it easier to compare one patient’s results to those of many others without having to manually sift through all of those disparate records. Cloud computing provides elastic scaling, so hospitals can use machine learning and AI to crunch their valuable patient data without buying dozens of new servers and racks. Taken together, these technological advances are making it easier to find the signal in the EMR noise.
As new data pours in and hospitals are able to harness advances from big data processing, a new focus on user experience should lead to intuitive tools that will help oncologists pull knowledge not only from each patient’s record, but also from a healthcare system’s global collection of records. Such tools could also greatly reduce the amount of time physicians spend on data entry.
Visualizing the patient’s healthcare journey, for example, could be transformed from a frustrating slog through the EMR to an at-a-glance, dashboard-style representation of the patient’s previous interactions with physicians, response to treatment, cancer progression, test results, and more. In addition, digital tools would make it easy for oncologists to see treatment response and other outcome information for all patients in the system whose cases are comparable to the patient in question — as well as to query the data and get results quickly. Imagine making clinical treatment decisions for a patient with the full context of how similar patients responded to the treatment options, including patients you might never have seen but who belong to your hospital’s population.
This kind of population-level data could help hospital systems quickly home in on new procedures that are showing positive results or flag new treatments that may have unexpected side effects. In addition, it could be used to spot patients who match clinical trial opportunities, or even to design clinical trials based on the specific needs of the people treated at a healthcare system. Building cohorts through digital tools would streamline the process of enrolling and launching new trials.
Digital tools to help oncologists derive more useful intelligence from disparate patient records are already being tested at many hospital systems, with early results indicating that these tools will make a real difference for improving both patient care and the user experience for physicians. Healthcare providers who have not had the opportunity to try these kinds of tools can reach out to their hospital administrators to arrange demonstrations or trials, or learn more about the array of digital tool options at key conferences and other venues. As more and more systems adopt digital health tools, oncologists can look forward to benefits in visualizing the patient journey, rapidly building cohorts for clinical trials, and other advantages from applying data mining tools to EMR platforms.
Head over to our Oncology page to read more about our offerings and how you can partner with us.
You can also use the form below to connect directly with our Oncology team.