The digital age of healthcare presents endless opportunities for providers to improve care, but all this information is useless if providers can't use the data to drive clinical decisions.
Natalie Pageler, MD, CMIO of Stanford (Calif.) Children's Health and Lucile Packard Children's Hospital Stanford, highlights how her practice has employed technology to give providers the right information at the right time.
Question: What makes analytics actionable to your physicians?
Dr. Natalie Pageler: The key to making analytics actionable to physicians is turning large amounts of data into meaningful information, and presenting that to providers at the right time and place in their workflow. Much of our advanced analytics work has focused specifically on embedding actionable information directly into clinical decision support tools to help physicians and other clinicians make the right decision at the right time.
For example, we developed a decision support tool called GluVue to help our diabetes doctors act on large amounts of patient-generated data that was collected from home glucose monitoring devices. The data itself isn't useful because it's so overwhelming, but using an EMR-integrated web-based visualization tool presents it in a meaningful way for the management of individual patients, and analytics reports support the management of our population of diabetic patients.
Q: Do you have programs using analytics to improve clinical decision making?
NP: As I mentioned above, we have focused many of our advanced analytics projects specifically on improving clinical decision making by embedding the information derived from advanced analytics directly into clinical decision support tools.
Another area where we have recently made progress is personalized dosing, which involves getting the right dose to the right patient. In pediatrics, this is especially challenging as there are large variations in dosing across difference ages. To help address this challenge, one of our pediatricians, Adam Frymoyer, MD, has collaborated with a Bay Area start-up company, InsightRx, which provides a personalized dosing platform for neonates, infants and children for several narrow therapeutic index drugs including vancomycin. The approach leverages patient-specific clinical and laboratory data within our EMR including drug concentration history to individualize the vancomycin dosing strategy for each patient (using advanced mathematical modeling and simulation techniques). The platform delivers this technology in the form of an easy-to-use clinical decision support tool that has been integrated within our EMR to help guide vancomycin treatment decision-making at the point of care. By personalizing the vancomycin dose for each patient, the treatment benefit can be maximized and toxicity minimized.
Q: What are some of the challenges using analytics?
NP: One of the major challenges in analytics is making meaning of the rapidly growing amounts of health data made accessible from many different sources. These sources range from patient-generated health data via home monitoring devices to massive amounts of genetic sequencing data coming from the Stanford Clinical Genomics service.
Q: What has been the progress you've made?
NP: We've made tremendous progress by identifying specific information gaps or areas where the amount of data has been overwhelming, and then partnering with data scientists across Stanford University and around the Bay Area to derive information that directly benefits our children and families.
In addition to the examples above, one of our recently graduated Clinical Informatics Fellows, Veena Goel, MD, partnered with Sarah Poole, a PhD student in the Stanford Biomedical Informatics Program, to evaluate several years' worth of pediatric vital sign data and develop more appropriate pediatric alarm limits to improve monitoring of children in our hospital. Additionally, David Scheinker, PhD, a data scientist in our hospital has been leading collaborations with Stanford School of Engineering students to help our pediatric surgeons more effectively predict surgical times and optimize surgical scheduling. This can be incredibly important to patients and families because the information helps decrease wait times before surgery, and helps provide families with more accurate information about what to expect throughout the process.