Bringing Population Health to the Point of Care
A data scientist wants everything to be data-driven. She will see data, how to learn something from that data, and how to iterate on it. I believe by making medicine data-driven and allowing every stakeholder to be empowered by the insights gleaned from the data, we can significantly improve outcomes for patients.
Lack of Data → Wrong Decisions
Doctors are at the focal point of making decisions day-in and day-out that can save and extend patients’ lives. When the wrong decisions are made, it is often because one of four things happen:
- The doctor doesn’t have the right context about the patient — their unique risks and propensities for disease;
- She doesn’t know what the most effective and evidence-based medical treatments are, given the unique features noted above;
- She doesn’t have the right mapping between context (1) and medical treatment (2);
- With significant workload constraints, there is no time to think through and implement the most effective patient management.
As the penetration of wearables increases and there are more healthcare applications that understand what the patient does outside the clinic, the problem #1 will become easier to tackle. Improved care coordination and workflow innovations can help significantly with the problem #4. The second and third items require bringing the power of medical science to bear upon each patient’s optimal treatment at the point of care.
Let’s take a look at diabetes, a direct contributor to increasing one’s 10-year risk of a cardiovascular event. Treating cardiovascular disease is ranked the highest, most expensive condition to treat in the United States, with a spend of $231 billion in 2013 with almost 60% coming from inpatient care.
Big Data → Small Data
When looking at the diabetes population, we can extract a significant amount of learnings from population health management to strive to keep patients out of the inpatient setting. By compiling all clinical data, we are able to create a data foundation to increase patient outcomes through:
- Identifying patients who are not up-to-date on tests for A1c, lipids, blood pressure and more;
- Ranking patients by number of deficits to prioritize outreach efforts;
- Measuring success of diabetes management interventions;
- Recognizing social determinants and behavioral components.
For an effective course of using population health, it all boils down to using data well. If we do, we can enhance the solution for all 4 problems with a combination of platforms.
CLINT for HealthPals solves the problems #2 and #3 above (and works with solutions that solve the first and last problems) by allowing doctors to have best practices and knowledge at their fingertips at the point of care. Only when we empower doctors with this methodology we will be able to persistently improve — improve how we practice, make new clinical discoveries, and save patient lives.