The Limits of Big Data
This morning in a press release Lilly (NYSE: LLY) stated;
“In an effort to improve health outcomes of people with type 2 diabetes, Geisinger and Boehringer Ingelheim, on behalf of its diabetes alliance with Eli Lilly and Company (NYSE: LLY), today announced a major collaboration to develop a risk-prediction model for three critical health outcomes commonly associated with type 2 diabetes that have long-term impact and cost-of-care implications: cardiovascular death, kidney failure and hospitalization for heart failure.
The new model will allow healthcare professionals to predict which adults with type 2 diabetes are most at risk for developing these serious – and costly – health consequences. The model will be created using Geisinger de-identified electronic health record data (i.e., demographics, vital signs, medical history, current medications and laboratory tests). Ultimately, a successful model should allow for the development of more precise treatment pathways for people with type 2 diabetes; pathways that align with quality guidelines aiming to improve patient outcomes, quality and total cost of care.”
Lilly is not the first nor likely will they be the last company to use big data to help improve outcomes for patients with diabetes. Although we are encouraged that Lilly is not limiting their efforts to just insulin using patients and is looking at ALL patients with diabetes, this is refreshing. Yet even with these efforts there is one undeniable fact that stands between big data and better patient outcomes; the patient.
Ask any physician that treats patients with diabetes what’s the biggest obstacle between patients and better outcomes and nine times out of ten they’ll point to medication adherence. This as we like to say is not news. For years even before the advent of big data we have known this to be true. Yes, we can always use better drugs and devices but the reality is the drugs and devices we have today work just fine when they are used.
This is one reason we have been so high on the Intarcia exenatide micro-pump as it solves the biggest obstacle standing between a patient and better outcomes. Is the system perfect, no, as there is no such thing. But when it comes to medication adherence few systems can match what Intarcia has developed. Once implanted in the body that’s it the patient goes about living their life.
Another approach to improving medication adherence are all the way cool whiz bang cloud enabled systems that come with patient coaching. The belief here is that these systems will transform data into patient action and therefore they will be more compliant with their prescribed therapy regimen. Secondarily the possibility exists that data analytics will discover a change is necessary in the patient’s therapy which will be relayed to their physician who in turn will make this change.
As much as we see the promise of interconnected diabetes management (IDM) we also see it’s limitations. As when its all said and done ultimately, it’s the patient who must follow through and follow the advice they have been given. This is in sharp contrast to the Intarcia system which once implanted the patient does nothing. This the same reason GLP-1 therapy continues to grow. When Byetta arrived, it was taken twice a day, which followed by Victoza which is once a day and now Trulicity and Bydureon which are taken once weekly.
The message here is clear the fewer times a patient needs to take their meds the better. This is the same reason every insulin company seeks to develop an ultra-long-acting insulin one taken every three days rather than taken every day.
Crazy as this is going to sound what most companies ignore; and this goes for drug, device and IDM companies, is the patient who uses the drugs, devices and IDM systems. They live under the false impression that these patients like being told what to do and when to do it. Simply put they are asking the patient to fit their drug and device into their lives when they should be asking how they can better fit their drugs and devices into the patient’s life.
Allow us for a moment to share some examples of how this can work. Back when Bydureon was in clinical trials one doctor we spoke with noted that he loved the drug for one simple reason, patients are creatures of habit. As he stated people do something once a week whether it’s doing the laundry or making a bank deposit. These activities have become part of their lives a habit. He just told his patients to align their injection of the drug to the same day they performed this activity.
The Libre is another example. Even though the device is not a full blown CGM with alarms and alerts it does not require calibration. Once on the patient is in control over when and how they use it. The lesson with the Libre is clear the less interaction the patient has with a device the better.
This to us is the biggest obstacle with the current crop of IDM systems instead of helping patients interact less with their diabetes management they are forcing the patient to interact more with their diabetes management. This is the exact opposite of what most patients want. Most patients want simplicity, less things to do instead of more things to do even when they are getting good advice. These patients want to think less about their diabetes when these systems make them think more about their diabetes.
It’s difficult to explain to someone who does not have diabetes the mental/emotional toll managing diabetes takes on a patient. Even those of us who are super engaged with our diabetes management and practice solid diabetes management experience this. Managing diabetes is hard work and at times very stressful. As we see it the problem is even worse for non-insulin patients as they do not feel the impact of say forgetting to take a pill. This impact will eventually rear its ugly head but the fact is these patients will not feel physical discomfort or pain until its too late.
This is why we have been hammering away on changing the patient’s experience with their diabetes management. How we need different motivational tools or incentives which more closely align to what’s important to the patient. The sad reality is the tools we have been using just don’t work yet everyone keeps trying to use them.
Like IDM this push to use big data has the potential to yield fruit but only if the patient cooperates. Patients cannot be forced to cooperate nor can they be guilted into cooperating. Patients will only cooperate when they see a benefit for doing so and that benefit is not something intangible like better outcomes. Patients cannot see, feel or touch better outcomes.
This may seem like common sense yet this very simple concept continues to be ignored.