Making treatment decisions is a tough choice for oncologists: a therapy might save or significantly extend one patient’s life but not deliver the desired outcome, for another patient. Reaching this decision involves weighing a variety of data – from clinical trials to the patient’s medical background – and with the advent of more personalized oncology, the sheer volume of data that needs to be considered is growing exponentially.
When we speak to oncologists – the consistent feedback we hear is the complexity of information is increasing and we need support to access it faster, more efficiently and in a more targeted manner.
Consider how precision oncology will fundamentally change the way cancer patients are treated.
First, instead of looking at the organ of the cancer’s origin only, doctors are going to pay more attention to the patient’s genomic characteristics and the medical history when deciding on the most effective cancer therapy. With next-generation sequencing technologies now broadly available, it is easier than ever before to understand the genomic variants of a cancer. We expect many new therapies based on the molecular profile of a tumor to come to the market in the next three to five years; for some indications it could be two or three times as many therapies compared to today.
<<< Start >>>
<<< End >>>
All the “new” data that is becoming available is driving therapeutic decisions in oncology
It is therefore legitimate to ask: Could oncologists be overwhelmed by the amount of data available to them? The short answer is: if limited to only today’s standard tools and practices, if they aren’t already, they will be very soon.
In the near future, oncologists will need to embrace digital solutions, such as Clinical Decision Support (CDS) systems, to manage the complexity. Based on algorithms and extensive computing power, CDS tools can structure and filter clinical data to help physicians make more informed treatment decisions faster. However, according to our recent research with 130 oncologists from the US and Europe, only one fifth of oncologists routinely use CDS tools today. We expect that number to significantly increase and CDS to become a standard tool for tumor diagnosis in the coming years. Furthermore, we believe that these tools will contribute to the breakthrough of precision oncology as they help physicians choose individual therapies for patients over standard treatments and do so with a reduced margin of error.
The main objective of CDS tools is to structure and filter information so the physician will only have to analyze data relevant to a specific case. Depending on the task – either diagnosis or assisting in treatment decisions – the way the CDS tool supports the physician is different. In the first case, CDS are significantly reducing the risk of human error by “spotting” things even the most experienced oncologists might overlook or at least need a second opinion for.
When it comes to deciding on the right treatment, CDS is not replacing the physician’s authority but rather provides and classifies relevant information that allows the specialist to make an even better and more informed decision. For example, CDS tools display the appropriate clinical evidence. As Dr. Diana Caragacianu, a medical director at the Breast Center at Milford Regional, points out: “these recommendations solicit our thoughts and I feel like we can be better doctors and can deliver better, more precise care.”
This feature becomes even more important as the number of possible therapy options increases once oncologists start considering genomic profiling for their patients. Dr. Nisha Unni, an assistant professor at UT Southwestern Medical Center, argues that, as treatment options are “going to get more granular and more detailed, it’s good to have a lot of treatment support.” This allows oncologists to use nonstandard treatments with more confidence or focus on findings of a clinical trial for a particular sub-population.
So why isn’t it used more now?
The complaint we’ve heard from physicians it that too much useless data is thrown at them and that CDS tools aren’t always very good in understanding the context in which decisions need to be made by physicians. Dr. Joshua Sabari, at Perlmutter Cancer Center at NYU Langone, illustrates this perfectly: “If you have a patient who's a smoker, I get a pop up about, Did you screen them for lung cancer?‘ I am a lung cancer doctor. All my patients have lung cancer.”
One-size-fits-all CDS tools simply ask too many questions or require too much effort by the physicians to get the user experience right. Instead of reducing complexity and saving time for the physician, one-size-fits-all CDS tools have the opposite effect in reality: They require physicians to filter much more information themselves as they need to carefully consider which click is important and which is not. The more the physician has a role in triggering the right information, the less likely the physician will use the tool – it needs to happen in the background.
But it will -- with improved algorithms and more real-world data being fed into these tools, CDS will become even more accurate in the future. Consequently, CDS tools will become the standard application in diagnostics. They are more likely to get the diagnosis 100 percent right compared to even the most experienced oncologist who might only be at 99 percent. CDS tools need to improve by requiring less front-end physician input than they do now.
To be clear, CDS systems are in the end just tools. The most effective therapy might not always be the best treatment option for the patient. Quality of life, for example, is a dimension that is hardly considered by algorithms although it is a very important aspect for patients who are going through a cancer therapy. It is the patient and the doctor deciding together what’s the right approach in an individual situation. But arming both the patient and physician with the best possible information to help make that decision is crucial.
To learn more about what oncologists say they need, read our full research report The Future Is now: How to Drive Precision Oncology