By Sara Russell Rodriguez, MSN, MPH, RN, Vice President Ė Integrated Care Management at Advocate Aurora Health
In population health, predictive analytics refers to the process of analyzing large amounts of aggregated health care data, in order to draw important insights and extract information to improve health and reduce costs. Effective population health management strategies require an in-depth look at a populationís current health ó including risk factors and patterns of care ó as well as tracking health care outcomes over time.
On a population-wide level, predictive analytics can help trim costs by accurately predicting which patients are at higher risk for poor outcomes, then devoting additional resources to those patients through early interventions. Data analysis allows providers to develop personalized risk profiles and identify cases where additional education, coaching, disease management and clinical integration have been shown to improve patient health.
The success of population health requires engaging patients in their own care and using a collaborative team approach. Hereís an example to illustrate the point:
Mrs. Smith (not her real name) is a 63-year-old type 2 diabetic. She is overweight and suffers from high blood pressure and high cholesterol. Her electronic medical record (EMR) shows she has poor control of her blood glucose levels, putting her at high risk for serious diabetic complications and future hospitalizations.
How can population health efforts using predictive analytics, help this patient stay well and potentially reduce health care costs she most likely will incur?
Many factors affect health care outcomes, including the patientís own health habits and lifestyle behaviors, as well as socio-economic factors (called social determinants of health). Research shows that diabetes is best controlled through medication adherence, diet, exercise and patient education. Health counseling high-risk individuals such as Mrs. Smith can greatly reduce health costs and
improve patientsí health.
Managing complex cases, where patients often have multiple medical conditions, requires collaboration among:
- primary care providers
- nurse navigators
- behavioral health professionals
- social workers
- complex case managers
- other medical specialists and subspecialists
In cases like Mrs. Smithís, proactive health interventions are proven to be cost-effective by reducing:
- disease complications
- frequency of emergency room visits
- number of inpatient hospital admissions
- 30-day readmission rates to hospital
- average length of hospital stays
Digging through data to developing best practices
Healthcare providers optimize their population health efforts through data mining
. This is key in developing best practices and standardizing the most effective medical treatments for patients. How is this accomplished? Data mining uses various sources of data: insurance claims, pharmacy claims and patientsí electronic health records.
Insurance claims data
Insurance claims data provides valuable information (e.g., diagnosis codes and dates of specific visits to a clinic or hospital, and the costs of those services). This helps team members understand who they are treating and the health issues the patient is facing. This information can be useful to employers, too. Without violating patient privacy laws, employers may want to learn if their employees (as a group) are using the hospital emergency room more than average, or if employees are using the preventive care benefits available to them.
Pharmacy claims data
Analyzing pharmacy claims can identify patients who arenít following their prescribed plan of care. Why didnít Mrs. Smith (in the example above) fill her prescriptions last month? The population health team can contact her to better understand her barriers to care. Analyzing pharmacy claims data allows the team to monitor how well patients are managing their conditions and adhering to their prescribed medication regimen.
Electronic health records data
Data mining electronic health records (EHRs) offers direct insights into clinical findings. It provides essential patient information that may trim costs by reducing unnecessary care or duplicative diagnostic testing. The EMR includes:
HR leaders: Understand whatís working (or not working)
- current health status,
- lab and diagnostic testing results,
- health history, and
- many other useful metrics.
Predictive analytics can effectively identify high-risk patients and uncover which practices are most effective in improving the health of a patient population. HR leaders
should work with their healthcare partners and payers to understand their own population health data. Data analysis provides a more complete picture of the health of an employee population. This data can provide key insights on whatís working well (and whatís not) ó hopefully offering new approaches to ensure that your workforce remains as healthy as possible ó while maximizing your health care investment in the process.