When data isn’t working; it’s not data, but misalignment
One recent case exemplifies the challenges facing healthcare providers trying to manage patients based on haphazard communication of data.
I recently commented on a post on LinkedIn where the care team missed an order to get a patient on hospice care. In a gut-wrenching video, the poster (Grace Cordovano) described how a metastatic cancer patient who had already lost everything lost the opportunity to die in peace because a hospice order wasn't entered.
One could blame the physician or the office for missing the order. Or theoretically, you could look to the patient or family for not noticing that the patient instructions had content about what to do to get hospice care. Or the nurse could or should have emphasized such an important item in the patient's long-term care plan.
All these have a certain level of validity. However, any such validity exists only in isolation. The problem is that we have too many of these items in our care teams' and patients' to-do list. It is a veritable whack-a-mole game we play to keep from being overwhelmed.
The challenges
And we are all getting overwhelmed, not only with clinical tasks but also with such things as documentation, billing, dealing with regulations and the steady stream of similar tasks with which we are constantly being hammered. The sum total of these tasks is overwhelming and it is impossible to keep up with the demand. It is an epidemic of "be more careful."
As a metaphor, can you imagine driving down a highway and, in addition to paying attention to the brake, accelerator and steering, you had to deal with text messages, respond to work emails and deal with your bickering children in the back seat? We wouldn't be surprised if this driver ended up in a severe collision. Likewise, we shouldn't be surprised by the consequences of the system we put into place for this cancer patient.
What can we do to address this and so many other similar situations?
We need to make the obvious, clear next step obvious and clear. We can't rely on coaxing (or berating) people to be more careful. We must assume people will miss the next step and create mechanisms that can accommodate our human tendency to be distracted.
How do we make such systems and how are they different from what we have now?
Currently, we largely approach managing our patients by individual encounters. In the case of this unfortunate patient, the decision to consider hospice was dependent on the patient being in front of the physician. Then, the system was dependent on the physician entering an order. Then, we rely on the nurse reviewing key elements of instructions with the patient. Then, we rely on the patient to remember anything the patient needs to do in follow up.
There are also any number of other stakeholders that are involved in this complex process, such as case managers and visiting nurses. For everything to go right, we need almost the opposite of Reason's Swiss cheese model of high reliability. For everything to go right, all the holes must match up exactly.
The needed solution
How can we address this? In this case, I would suggest that, if good population health structures and processes had been in place, it would have been more difficult for this patient to fall through the cracks.
I would suggest this would start with a registry of metastatic cancer patients. There would be a sequence of metrics that could be created and tracked in a visualization that showed if patients are being appropriately managed.
In the case of this and similar patients, we want to identify patients with these characteristics – does the patient have advanced cancer? Should we consider hospice? Has someone asked the patient about hospice? Has the patient agreed to hospice? Has the order been placed? Has hospice contacted the patient?
As simple as these questions seem, creating them as discrete, accessible data elements is not simple. Therefore, creating processes to create an infrastructure to house these data would be a prerequisite. This is hard work, requiring domain expertise and organizational and technical skills.
After these data are in place and available, we would need to organize the data and visualize the elements in a way that makes sense to the staff managing these patients.
Then, we have to create actions based on insights garnered from the visualizations. Which advanced cancer patient should we consider hospice? Which of these patients has not had a conversation about hospice? And in the situation that occurred with this patient, which of these patients is missing an order?
Of course all this does require an alignment of incentives. I recently had a conversation with a healthcare business intelligence analyst. She had built some beautiful population health dashboards for her physicians. They ignored them because they didn't have time to address care gaps because doing so didn't bring them additional revenue and they were already overwhelmed with fee-for-service and encounter-based work.
I'm not saying all this is easy, but staying with our current system is worse. Our economy suffers, our providers suffer and, most importantly, our patients suffer.
John Lee, MD, has been a clinical informaticist since 2006 and has been Chief Medical Information officer at two health systems. In 2019, he was honored with the prestigious HIMSS Physician Executive of the Year award. He is a firm believer that the key to solving the myriad problems we experience in our healthcare system is efficient and transparent delivery of information.