With the healthcare landscape changing from fee-for-service to fee-for-value models, healthcare provider systems (hospitals, clinics, independent physician associations, etc.) are now, more than ever, under pressure to effectively manage the health and cost outcomes of their given populations. Under such models, providers are not only providing healthcare service to the patients, but they are also sharing in the financial risk and reward of patient costs. To effectively become a value-based organization, providers today are adopting a process broadly termed "population health."
The "population health" process usually starts with identifying key segments of a population that face certain risks of adverse health outcomes and thereby high cost — a step known as "risk stratification." Once risk is stratified, appropriate patient intervention programs are employed to improve: access to health, targeted encounters with providers and continuous monitoring of patient risk. This leads to lower emergency room visits, better clinical outcomes (such as properly managed blood glucose levels for diabetics) and lower financial cost.
There are many proven methods of risk-stratification to assign patients to low-, medium- or high-risk groups. For example, the adjusted clinical groups method examines patient diagnoses, and the elder risk assessment method assigns risk based on patient demographics. In today’s market, we observe many proprietary methods of risk stratification developed by various provider systems. The variables used in risk stratification can be classified into the following categories:
- Clinical: Data from electronic medical records (EMRs), patient vitals, laboratory data, etc.
- Administrative: Usually patient claims that track diagnosis and procedures already conducted
- Socio-Economic: Patients’ social situations, family and friend support systems, language preference, community involvement, the degree of influence that out-of-pocket expenses could have on the patient’s well-being, etc.
- Lifestyle: Health and activity tracking devices such as Fitbit, Apple Watch, etc., which carry critical daily lifestyle data about a patient
- Gather patients’ social network geographic information
Most EMR systems already contain patient address information, but they often lack information about the patients’ social network. The following types of data should be collected and refreshed on an annual basis:
- Distance to closest primary care clinic, both straight-line and network-distance;
- Distance to closest primary care provider, both straight-line and network-distance;
- Spatial density of medical resources in a given area, especially primary care services;
- Access to vehicle transportation, either on the patient's own or through a family member; and
- Proximity to public transportation.
- Conduct “spatial access” risk stratification Using a geographic information system (GIS), assign relative risk to each patient based on each of the components listed above, then create a composite risk based on all of the attributes.
- Represent population risk stratification visually via mapping Examine which areas of a provider’s service areas are prone to having individuals with high risk; look for clusters of high- or low-risk patients in doctor deserts. Viewing individual or aggregate risk through mapping would offer analysts and decision makers a comprehensive view of what types of risk are occurring in their service area.
- Strategize how to implement interventions based on locations of high-risk patients If clusters of high-risk patients exist in a certain area, begin to strategize about what kinds of interventions may alleviate the problem. Interventions may include the placement of new primary or specialty care clinics. Because creating clinics can be challenging, increased use of mobile provider teams can be an alternate solution. Lastly, a combination of telemedicine and mobile medicine should be assessed for the right mix of care for doctor deserts and lack of physical clinics.
