Skip to main content

Stanford Children's CMIO Talks Home-grown Clinical Decision Support

Analysis  |  By Alexandra Wilson Pecci  
   July 25, 2017

Innovating on top of the EHR enables clinicians to get information about a patient, create clinical decision support, and integrate the data with the workflow, says Natalie M. Pageler, MD.

Clinical decision support is getting more personalized and nuanced all the time.

HealthLeaders recently talked with Natalie M. Pageler, MD, chief medical information officer at Stanford Children's Health and clinical associate professor of pediatric critical care and Stanford University, about some of the tools that the hospital is developing.

This is the first of two parts. The transcript below has been lightly edited.

HLM: Tell us a bit about your role.

Pageler: I am the liaison between the operational clinical side and the information services department. I oversee a large team of clinical informaticists, which is comprised of physicians, bedside nurses, and respiratory therapists. I also oversee the EHR training team.

HLM: You had a lot to do with creating a clinical decision support tool that is personalized for every patient. Can you describe it?

Pageler: We've had an approach that's involving web-based EHR clinical decision support tools. We really believe there is a huge opportunity to innovate on top of the EHR, and that by developing tools that are web-based, that have links with the EHR, we can get information about the patient, create clinical decision support, and then show that back in the context of the EHR so that it's integrated into the workflow.

A few specific examples: One is a new tool that we're creating called Premie BiliRecs that's being led by one of our neonatologists and clinical informatics physicians, Jonathan Palma, MD.

The tool takes information from the EHR about a premature baby's date of birth and their bilirubin levels, then processes that and returns guidance on how to treat that bilirubin level in that neonate. Bilirubin can be very toxic to neonates, and especially premature babies.

[The tool is] actually live, and one of the unique things about [it] is that we are taking information from that tool and constantly learning about the response to the intervention and seeing how those patients do, because there are not completely standardized guidelines for premature babies and how to treat bilirubin. This tool is actually helping us to learn from our experience and feeding that information back into the tool provides better guidance.

HLM: Can you give us another example of a clinical decision support tool in use?

Pageler: Our hospital is located in Silicon Valley, which is a very fortunate position to be in. As part of the Stanford University campus, we have the opportunity to partner with both students in the biomedical informatic program, computer science programs, as well as multiple local startups.

Another type of advanced clinical decision support tool, again [one that's] web-based, and HER-integrated, is a product that we have been collaborating on with a small startup company called InsightRX.

Again, we're trying to create really personalized clinical decision support. In this case, we pass information—deidentified information—from the EHR about pediatric patients' age, weight, and some lab values, and then this tool provides specific dosing guidance for that particular patient.

It's very important because dosing in pediatrics is incredibly challenging with the varying weights and varying metabolisms of our patients.

For example, vancomycin is an antibiotic that can have significant toxicity in pediatric patients. This clinical decision tool has been implemented, and right now [is] specifically focused on vancomycin dosing.

HLM: These are super specific. Not only to the patient, but to what they're treating. Is that the future of clinical decision support in your opinion, or is there space for it to be broader?

Pageler: I think there are both options. We have very broad clinical decision support tools like our general medication dosing guidelines that span across all of our medications.

But I think the opportunities as we go forward are really to create more and more specific clinical decision support based on details about that particular patient, really getting to that idea of precision medicine.

What I think we're finding as we go forward, is that not all guidance applies equally to all patients, especially in pediatrics.

So the more precise we can get with our decision support, the better outcomes we'll have. As we see developments in genomics for example, we will continue to get more and more precise guidance for caring for our patients.

HLM: How do these tools and the data from these tools link into the patient's broader EHR? Would another physician be able to see the information generated?

Pageler: It depends, but the idea is that we want to facilitate our learning health system.

So we are getting guidance from some of these clinical decision support tools, acting on that guidance, learning how that information affected that patient, and then feeding that back into our overall learning cycle and continuing to enhance that clinical decision support.

For different tools, we have different ways of recording information into the EHR or into a research database depending on what that specific project is.

HLM: Could you give an example of one of those?

Pageler: The two examples I gave you are exactly that. So for Premie BiliRecs, it is provider guidance based on best practice right now, but all of the information is also being collected in a research database that is then being evaluated to continually update our decision support tools.

For the InsightRX dosing tool, again information is collected by the tool to help continually update it.

Another example is we used Apple HealthKit technology to upload continuous glucose monitoring for our diabetic patients, and so their glucose data goes directly into our EMR.

That is then sent to another web-based EMR-integrated clinical decision support tool, GluVue, which then displays that data in a very easy-to-digest format for the endocrinologist so they can make decisions using that data. And all of that data is recorded into the EHR.

Alexandra Wilson Pecci is an editor for HealthLeaders.


Get the latest on healthcare leadership in your inbox.